The modification of a commercial nanofiltration (NF) membrane (TFC‐S) with shape‐persistent dendritic molecules is reported. Amphiphilic aromatic polyamide dendrimers (G1–G3) are synthesized via a divergent approach and used for membrane active layer modification by direct percolation. The permeate samples collected from the percolation experiments are analyzed by UV‐visible spectroscopy to monitor the influence of dendrimer generations on percolation behavior and active layer modification. Further characterization of modified membranes by Rutherford backscattering spectrometry and atomic force microscopy techniques reveals a relatively low‐level accumulation of dendrimers inside the original TFC‐S NF membrane active layer and subsequent formation of a coating of pure aramide dendrimers on top of the active layer. A PES‐PVA ultrafiltration membrane is used as a control membrane support (without an NF active layer) showing that structural compatibility between the dendrimers and support plays an important role in the membrane modification process. The performance of the modified TFC‐S membrane is evaluated on the basis of the rejection abilities for a variety of water contaminants having a range of molecular size and chemistry. As the water flux is inversely proportional to the thickness of the active layer, the amount of dendrimers deposited for specific contaminants are optimized to improve the solute rejection while maintaining high water flux.

The fully aromatic polyamide active layer of a commercial nanofiltration membrane was modified with three generations (G1, G2, and G3) of aramide dendrimers, all with oligoethylene glycol chains on their peripheries. Permeation experiments revealed that the rejection of Rhodamine WT, used as a surrogate for organic contaminants, improved 1-2 orders of magnitude for membranes modified with G2 and G3 dendrimers at loadings of 0.7-3.5 μg/cm(2) (dendrimer layer thicknesses of ~1-6 nm) compared to the performance of unmodified membranes. In contrast, the corresponding water permeability of dendrimer-modified membranes decreased by only ~30%. Although an enhancement in the rejection of H(3)AsO(3), NaCl, and BaCl(2) was also observed for dendritic membranes, the effect was less pronounced than that for rhodamine WT. Characterization of membranes modified with 3.5 μg/cm(2) dendrimers G2 and G3 by Rutherford backscattering spectrometry with the aid of heavy ion probes (Ag(+) and Ba(2+)) revealed that accessibility of the larger Ba(2+) probe to carboxylate groups on the active layer decreased for the membranes modified with dendrimers.

A first generation of amine terminated aramide dendrimers (G1-NH2) was covalently attached to the polyamide (PA) active layer of a commercially available nanofiltration (NF) membrane. Amide bonds between G1-NH2 and PA free carboxylic groups were formed by activation of the carboxylic groups with 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) or 2-chloro-1-methylpyridinium iodide (CMPI), followed by aminolysis. Dendrimer attachment was assessed by indirectly measuring the concentration of carboxylic groups and amine groups before and after membrane modification with RBS using yttrium and tungstate ions (Y(3+) and WO4(2-)) as ion probes. RBS analyses showed a decrease in the concentration of carboxylic groups and an increase in amine groups on the membrane active layer, consistent with dendrimers attaching covalently to the active layer. Permeation experiments with Rhodamine WT (R-WT) revealed that the water and solutes permeability decreased after modification with dendrimer G1-NH2. Water permeability of G1-NH2 modified membrane decreased by 16-19% using EDC combined with sulfo-N-hydroxysuccinimide (s-NHS), and by 17-33% using CMPI. The permeability of the electrolyte BaCl2 decreased by 54% after G1-NH2 modification using EDC/s-NHS and only by 20% using CMPI, the latter consistent with a weaker Donnan exclusion effect. The permeability of the larger solute R-WT decreased by 82% in modified G1-NH2 membranes when using EDC/s-NHS, and 64% for cross-linking reagent CMPI. Thus, the use of EDC/s-NHS was more favorable because it resulted in higher gains in solute rejection with lower losses in water permeability.