Pervaporation studies with polyaniline membranes and blends

Permeabilities (P), diffusion coefficients (D) and time lags (Theta) for wa ter, alcohols and carboxylic acids passed through undoped and doped polyani line membranes are reported. Permeabilities through undoped polyaniline are stable and linear, but both permeabilities and diffusion coefficients tend to be low. As an example, for water, P=0.111+/-0.05 g mm/m(2) h, with a di ffusion coefficient of D=5.0x 10(-10) cm(2)/s. Methanol has the highest per meability through undoped polyaniline (P=0.83+/-0.02g mm/m(2) h), despite i ts small diffusion coefficient (D=1.5x10(-10) cm(2)/s). Fully HCl-doped pol yaniline membranes have higher permeabilities (up to 5-6 g mm/m(2) h) with correspondingly larger diffusion coefficients (e.g. D-water>1.0x 10(-8) cm( 2)/s) and shorter time lags. Larger alcohols (e.g. 2-propanol) and carboxyl ic acids (e.g. acetic acid) show no permeation through doped polyaniline. T herefore, fully HCl-doped polyaniline membranes can be used to effectively separate water from acetic acid (alpha H2O approximate to 200). Unfortunate ly, when HCl-doped polyaniline is used as a pervaporadon membrane, the perm eation rates of the feed decline over time. For example, the permeability o f water declines from a high of similar to 2.5 g mm/m(2) h to a much lower average steady-state value of 0.3 g mm/m(2) h over 48 h. This effect has be en traced to the leaching of HCl dopants out of the polyaniline membrane an d into the feed solution. The problem of dopant leaching can be eliminated by using a polymer dopant in a blend with polyaniline. The permeation rates of ethanol/water feeds through blends of 10 wt.% polyamic acid/90 wt.% pol yaniline and 10 wt.% polyacrylic acid/90 wt.% polyaniline are linear and st able. This demonstrates that polymeric dopants eliminate the problem of dop ant leaching. Both blends possess permeabilities and selectivities in-betwe en undoped and doped polyaniline. (C) 2000 Elsevier Science B.V. All rights reserved.