Structure-property relationships for novel wholly aromatic polyamide-hydrazides containing various proportions of para-phenylene and meta-phenylene units III. Preparation and properties of semi-permeable membranes for water desalination by reverse osmosis separation performance

Flat sheet asymmetric reverse osmosis membranes were successfully prepared from N,N-dimethylacetamide (DMAc) solutions of a series of novel wholly aromatic polyamide-hydrazides that contained different amounts of para- and meta-phenylene rings. These polyamide-hydrazides were synthesized by a low temperature solution polycondensation reactions of either 4-amino-3-hydroxybenzhydrazide or 3-amino-4-hydroxybenzhydrazide with an equimolar amount of either terephthaloyl dichloride TCl], isophthaloyl dichloride ICl] or mixtures of various molar ratios of TCl and ICl in anhydrous DMAc as a solvent. All the polymers have the same structural formula except of the way of linking phenylene units inside the polymer chains. The content of para- to meta-phenylene moieties was varied within these polymers so that the changes in the latter were 10 mol% from polymer to polymer, starting from an overall content of 0-100 mol%. All the membranes were characterized for their salt rejection (%) and water permeability (cm³ cm⁻² day⁻¹) of 0.5 N aqueous sodium chloride feed solution at 3924 kPa operating pressure. The effects of polymers structural variations together with several processing parameters to achieve the best combination of high selectivity and permeability were studied. Effects of various processing parameters of the membranes on their transport properties were investigated by varying the temperature and period of the solvent evaporation of the cast membranes, coagulation temperature of the thermally treated membranes, annealing of the coagulated membranes, casting solution composition, membrane thickness and the operating pressure. During the thermal treatment step, the asymmetric structure of the membranes with a thin dense skin surface layer supported on a more porous layer was established. The former layer seems to be responsible for the separation performance. The results obtained showed that membrane performance was very much influenced by all of the examined processing variables and that membranes with considerably different properties could be obtained from the same polymer sample by using different processing parameters. Thus, the use of higher temperatures and longer exposure times in the protomembrane forming thermal treatment step would result in a membrane of lower solvent content and with a thicker skin layer and consequently led to higher salt rejection at lower water permeability. Most significantly, the membrane properties clearly depended on the polymer structure. Under identical processing condition, substitution para-phenylene rings for meta-phenylene ones within the polymer series resulted in an increase in salt rejection capability of the membranes. This may be attributed to an increase in their chain symmetry associated with increased molecular packing and rigidity through enhanced intermolecular hydrogen bonding. This produces a barrier with much smaller pores that would efficiently prevent the solute particles from penetration. Coagulation temperature controls the structure (porosity) of the membrane particularly its supported layer and consequently its water permeability. Moreover, annealing of the prepared membranes in deionized water at 100 °C was found essential for useful properties in the single-stage separation applications, which required optimum membrane selectivity. Upon annealing, the membrane shrinks resulting in reducing its pore size particularly in the skin layer and consequently improving the salt rejection. Addition of lithium chloride to the casting solution produced a membrane with increased porosity and improved water permeability. Salt rejection capability of the membranes is clearly affected by the applied pressure, reaching its maximum at nearly 4000 kPa. Furthermore, the water permeability is inversely proportional to the membrane thickness, while the salt rejection is not substantially influenced.