Polyamide-based membranes with structural homogeneity for ultrafast molecular sieving

Thin-film composite membranes formed by conventional interfacial polymerization generally suffer from the depth heterogeneity of the polyamide layer, i.e., nonuniformly distributed free volume pores, leading to the inefficient permselectivity. Here, we demonstrate a facile and versatile approach to tune the nanoscale homogeneity of polyamide-based thin-film composite membranes via inorganic salt-mediated interfacial polymerization process. Molecular dynamics simulations and various characterization techniques elucidate in detail the underlying molecular mechanism by which the salt addition confines and regulates the diffusion of amine monomers to the water-oil interface and thus tunes the nanoscale homogeneity of the polyamide layer. The resulting thin-film composite membranes with thin, smooth, dense, and structurally homogeneous polyamide layers demonstrate a permeance increment of ~20–435% and/or solute rejection enhancement of ~10–170% as well as improved antifouling property for efficient reverse/forward osmosis and nanofiltration separations. This work sheds light on the tunability of the polyamide layer homogeneity via salt-regulated interfacial polymerization process.


PALS characterization of as-fabricated TFC membranes
The detailed experiment procedure for PALS characterization is described below.
A variable mono energetic slow positron beam was operated in the range of 0-30 keV positron incident energy (equivalent to a mean depth of 0−10 μm) at room temperature under a vacuum of ~10 -8 torr. The positron source is of a 50 mCi of 22 Na radioisotope beam. Each membrane sample for PALS characterization was in area of 1*1 cm 2 . The as-fabricated membrane samples were measured by two positron annihilation techniques, including Doppler broadening of energy spectra (DBES) and positron annihilation lifetime (PAL) measurements. The DBES spectra were measured using a solid state HP Ge detector (EG & G Ortec) at a counting rate of approximately 3000 counts per second (cps). The total number of counts for each DBES was 1.0 million. The S parameter data from DBES was fitted by VEPFIT program. The PALS spectra were obtained by taking coincident events between two signals-the start signal detected by a multichannel plate from the secondary electrons and the stop signal discerned by a BaF 2 lifetime detector from the annihilation photons at a counting rate of ~200-300 cps. A PALS spectrum contains 2.0 million counts. The positron lifetimes (t) and intensities (I) were determined by PATFIT program, and the lifetime distribution was obtained by MELT analysis.

Supplementary Figures
Supplementary Figure 1. NaCl concentration in the washing water of the MPD@NaCl membrane. The error bars represent the standard deviation and were calculated on the basis of at least three data points measured from different samples. It can be seen that NaCl concentration in the washing water sharply decreases after the first day immersion and gradually slows down in the following days. It can be deduced that a large amount of salt removed in the first day is generally the salt trapped in the substrate membrane as well as between the substrate membrane and the PA layer. However, after the first-day washing, the salt trapped in the PA layer can be partially washed out with a longer washing duration (5 days) due to the complexion between the salt ions and the carbonyl groups of the PA network [15][16][17]23,24 . The increased free volume of NaCl-modified membrane is solely resulted from the removal of trapped NaCl salts. Instead, the increased free volume of NaHCO 3 -modified membrane is contributed by both the removal of trapped inorganic salts and generated CO 2 nanobubbles by the reaction between NaHCO 3 and IP byproduct HCl. However, the later factor is believed to outweigh the former factor. NaHCO 3 in the reaction zone would be consumed by the byproduct HCl generating CO 2 as shown in Supplementary Equation 1. It can be seen that 1 mole NaHCO 3 7 consumed will generate 1 mole NaCl and 1 mole CO 2 . Supplementary Figures 3b and   3d show that S values of NaHCO 3 -modified membranes also increase after washing, indicating the increased free volume resulted from the partial removal of generated NaCl. However, the trapped salt in the PA layer can be only partially removed according to the AAS result (< 3%). Meanwhile, all generated CO 2 can create nanobubbles in the PA layer. Therefore, the increased free volume of NaHCO 3 -modified membranes should be predominantly contributed by the nanobubbles of CO 2 generated.

1) TEA acted as a catalyst to accelerate the MPD-TMC reaction by neutralizing
HCl produced during amide formation.
2) The addition of CSA protected the microporous skin layer of the support membrane from annealing during curing. 2) BMIC as a phase transfer catalyst, not only improved the transfer mechanism across the interface, but also increased the diffusion rate. [3] SDS+CTAB PIP 1) The porosity of thin layers decreased in the presence of CTAB and SDS.
2) The extent of cross-linking was increased with the increasing molecular weight of cationic amine groups. 2) The ammonium salt with more and longer lipophilic alkyl groups showed the higher catalytic efficiency.
3) The larger ammonium salt complex would take over larger space in polymer, and result in a larger free volume. [6] Surfactant + amine salt Amine The salt effect was not stated. [7][8][9] Amine salt Amine The salt effect was not stated. [10] Surfactant + amine salt Amine The salt effect was not stated. [11] Tertiary amine salt Amine 1) Tertiary amine acted as both pore forming agent and catalyst (absorbing the IP reaction by-products). 2) Li + ions induce an increased interaction coefficient between TMC and TEOA, [15] 31 forming a dense skin layer.
3) Complexation between Li + ion and the carbonyl in TMC causes the hydrolysis of acid chloride groups of TMC, inducing a hydrophilic and loose surface layer.

CaCl 2 TEPA
1) The addition of CaCl 2 enhanced the interfacial tension and reduced the mass transfer of TEPA to the organic phase.
2) Higher CaCl 2 content induced an unevenly distributed interfacial polymerization reaction, resulting in the increased surface roughness. [16] TEA+NaCl (Nanofiber support) PIP 1) The addition of salt to the aqueous phase increased the interfacial tension, depressing the mass transfer of PIP and constructing a loose PA layer.
2) Further increase NaCl content induced the uneven distribution in nanoscale of the IP reaction, leading to the morphological structure transition from the crisscrossed ridge networks to crowed nodular arrays. 2) The density and size of NaCl crystals increase with the increasing NaCl concentration, resulting in unidirectional scattering distribution to interconnected dendritic distribution. [20] NaCl-modified membranes with and without washing by EDX analysis. It can be found that Na and/or Cl content in the NaHCO 3 and NaCl-modified FO membranes sharply decrease after washing.