Tuning polyamide membrane chemistry for enhanced desalination using Boc-protected ethylenediamine and its in situ Boc-deprotection

Polyamide thin-film composite (TFC) membranes formed by interfacial polymerization (IP) have replaced cellulose-based integrally skinned asymmetric membranes due to superior flux, salt rejection, and scalable fabrication. Despite successes, key challenges remain: the flux–rejection trade-off, chlorine resistance, antifouling, and lifespan. IP’s self-limiting kinetics, self-sealing behavior, and crosslinking (e.g., TMC as trifunctional monomer) make it a powerful platform to tune active-layer chemistry. This study addresses whether introducing a Boc-protected linear aliphatic diamine (EDA) into an MPD/TMC IP system, followed by in situ Boc deprotection, can enhance hydrophilicity, flux, salt rejection, and antifouling relative to a conventional MPD/TMC membrane. The hypothesis is that incorporation of EDA-Boc alters microstructure and chemistry, and subsequent acid deprotection generates surface ammonium groups that improve water transport and ion exclusion.

Two prevalent strategies exist to improve polyamide RO membranes: (1) new monomer chemistries in IP, and (2) post- or co-modification via additives. The former is less explored due to synthetic complexity; the latter has many reports. Additive approaches include m-phenylenediamine-5-sulfonic acid post-addition yielding ~1.4× flux increase at comparable NaCl rejection (Yong et al.); aqueous-phase regulation using linear aliphatic amine 1,3-diamino-2-hydroxypropane (DAHP) giving ~22% flux improvement with 96–98% rejection (Perera et al.); and using aliphatic basic amino acids as aqueous additives: L-arginine improved flux from 46 ± 0.5 to 54 ± 0.5 L m⁻² h⁻¹ and salt rejection from 96 ± 0.5% to 98 ± 0.5% (Chen et al.), and L-lysine offered ~18% flux enhancement without compromising rejection (Xu et al.). These works suggest that introducing flexible aliphatic chains and hydrophilic functionalities (amino, carboxyl, sulfonic acid) can increase permeability while maintaining high ion rejection, motivating the current EDA-Boc strategy.

Synthesis of EDA-Boc: Ethylenediamine (9.0 mL, 0.14 mol) and di-tert-butyl dicarbonate (Diboc, 2.6 g, 0.012 mol) were dissolved separately in DCM (100 mL and 5 mL). Boc solution was added dropwise to EDA at 0 °C with vigorous stirring, then stirred at 25 °C for 6 h. Reaction progress was monitored by TLC. Solvent was removed, crude washed with NaHCO3, water, and brine, and purified through silica. Structures matched prior 1H/13C NMR data. Support fabrication (NIPS): PSf/PET supports were prepared from a dope of PSf (18.0 g), PVP (2.0 g) in DMAc (80.0 g), stirred at 70 °C to clarity and degassed overnight. Cast on PET (Novatexx-2413) using 100 µm doctor blade; immediately immersed in DI water to form PSf/PET support; soaked in water 24 h before IP. Interfacial polymerization: Two aqueous amine solutions: (i) MPD 2.0 wt%; (ii) MPD 1.90 wt% + EDA-Boc 0.10 wt%. Organic phase: TMC 0.15 wt/vol% in n-hexane. Supports were immersed in amine solution for 10 min, excess removed by rubber roller, air-dried briefly, then contacted with TMC solution for 60 s. Membranes were rinsed with clean n-hexane and cured at 60 °C for 15 min to yield MPD-TMC (control) and MPD-TMC-EDA-Boc membranes. Boc deprotection: MPD-TMC-EDA-Boc membranes were treated with 100 mL aqueous HCl (20% v/v) for 6 h, then washed with DI water, producing MPD-TMC-EDA-Deboc membranes. Characterization: ATR-FTIR (Nicolet iS50) on free-standing active layers prepared under identical IP conditions; solid-state 13C CP/MAS NMR (Jeol-600 MHz) for MPD-TMC-EDA-Boc and -Deboc; SEM (Quattro SEM) for surface and cross-sections; EDX spectra and elemental mapping; AFM (Easy Scan 2, tapping mode) for roughness (10×10 µm scans); water contact angle (WCA) using goniometer (KRUSS DSA 20). Elemental CHNS/O analysis (Flash Smart). Bulk porosity via ε(%) = (Wwet − Wdry)/(V ρw) × 100. Degree of crosslinking from elemental O/N using standard MPD/TMC equations. Performance testing: Crossflow system (BONA-TYLG-19) with three membranes in parallel. Compaction at 30 bar to steady flux. Feed solutions: DI water; single salts (NaCl, MgCl2, Na2SO4, MgSO4) at 2 g/L; and real seawater nanofiltration permeate (TDS 33,700 ppm; provided by Saudi Water Authority). Tests at 20–30 bar as indicated. Salt rejection from conductivity: Rejection (%) = (1 − C2/C1) × 100. Flux J = V/(A t). Antifouling: 200 ppm bovine serum albumin (BSA) feed for 6 h at 20 bar; flux decline monitored. Cleaning by backwashing with DI water at 20 bar for 1 h; flux recovery calculated.

- Chemistry confirmation: ATR-FTIR of free-standing layers showed amide carbonyl (~1600 cm⁻¹) and N–H stretching (~3400 cm⁻¹). Upon Deboc, a broad N–H band indicated altered amine functionalities and higher hydrophilicity. Solid-state 13C NMR of EDA-Boc layer had aliphatic peaks at 19.7, 32.1, 38.1 ppm (Boc/EDA) and aromatic carbons 106.6–138.6 ppm; amide carbonyl at 155.3 ppm. After Deboc, Boc peaks disappeared; aliphatic signals shifted/merged toward aromatic region due to ammonium formation, while amide carbonyl remained, evidencing successful deprotection and ammonium ion generation. - Elemental composition and crosslinking (free-standing active layers, Table 1): • MPD-TMC: C 62.9%, O 19.9%, N 17.1%, Cl 0; O/N 1.16; crosslinking degree 77.8% (X=0.778). • MPD-TMC-EDA-Boc: C 67.9%, O 17.0%, N 15.0%, Cl 0; O/N 1.13; crosslinking 81.7% (X=0.817). • MPD-TMC-EDA-Deboc: C 68.8%, O 15.3%, N 13.7%, Cl 2.1%; O/N 1.12; crosslinking 83.0% (X=0.830). Increased C and reduced O consistent with Boc incorporation and loss of tert-butoxy groups after deprotection; Cl present as counterion to ammonium. - Morphology and thickness: Surface SEM showed MPD-TMC with ridge–valley network; EDA-Boc smoother fine layer; Deboc restored sharper ridges/deeper valleys due to removal of bulky Boc and oligomers. Cross-sections showed intact asymmetric structure (PET nonwoven/PSf UF support/polyamide skin). Active-layer thickness: MPD-TMC 145 nm; EDA-Boc 244 nm; EDA-Deboc 187 nm (thinner after Deboc, implying reduced mass-transfer resistance). - Porosity and roughness: Bulk porosity: MPD-TMC 9.6%; EDA-Boc 5.7%; EDA-Deboc 6.6%. AFM average roughness Ra: MPD-TMC 7.32 nm; EDA-Boc 6.27 nm; EDA-Deboc 7.87 nm. - Wettability (WCA): MPD-TMC 70°; EDA-Boc 68°; EDA-Deboc 55°, showing highest hydrophilicity post-Deboc due to surface ammonium groups. Roughness–wettability relationship indicates chemistry dominates for Deboc membrane. - Flux vs pressure (DI water): At 30 bar, fluxes were MPD-TMC 26.0 ± 0.2 LMH; EDA-Boc 33.0 ± 0.5 LMH; EDA-Deboc 37.0 ± 0.2 LMH. At 20 bar with DI water: MPD-TMC 22 ± 0.5 LMH; EDA-Boc 23 ± 0.5 LMH; EDA-Deboc 27 ± 0.5 LMH. - Salt rejection at 20 bar (2 g/L): NaCl rejection >93% for all; EDA-Boc 95 ± 0.5%; EDA-Deboc 98 ± 0.5%. Divalent salts >95% for all; EDA-Deboc highest (>98%), with MgCl2 99.0 ± 0.5%; MgSO4 98.5 ± 0.5%; Na2SO4 98.0 ± 0.5%. - Flux with salts at 20 bar (EDA-Deboc): NaCl 25 ± 0.5 LMH; Na2SO4 23 ± 0.5 LMH; MgCl2 23.5 ± 0.5 LMH; MgSO4 23 ± 0.5 LMH. Anion identity influenced flux (stronger anion affinity lowered flux). For seawater NF permeate (TDS 33,700 ppm): flux MPD-TMC 18.5 LMH; EDA-Boc 20 LMH; EDA-Deboc 23 LMH. - Real feed rejection: Seawater NF permeate TDS rejection: MPD-TMC 92%; EDA-Boc 94%; EDA-Deboc 97% at 20 bar. - NaCl desalination headline (20 bar, 2000 ppm): EDA-Deboc achieved 98 ± 0.5% rejection with 25 LMH flux; ~25% flux increase vs control. - Antifouling (200 ppm BSA, 6 h, 20 bar): Flux declines MPD-TMC 21 → 19 LMH (~10%); EDA-Boc 22 → 19 LMH (~11%); EDA-Deboc 27 → 24 LMH (~12%). After DI backwash (20 bar, 1 h), EDA-Deboc showed highest flux recovery 95.0 ± 0.5% (abstract notes 95 ± 1% vs control 93 ± 1%). - Comparative context: The MPD-TMC-EDA-Deboc membrane exhibits high pure-water flux (37 ± 0.2 LMH at 30 bar) and >98% monovalent rejection, competitive with or better than many reported RO polyamide membranes without the drawbacks of inorganic filler-based TFNs.

Introducing EDA-Boc into the MPD/TMC IP reaction alters active-layer growth, producing a smoother, thicker layer with slightly higher crosslinking and lower porosity. Subsequent in situ Boc deprotection with HCl removes bulky tert-butoxy groups and residual oligomers, thinning the active layer and increasing roughness while generating surface ammonium groups that substantially increase hydrophilicity (WCA 55°). The combined effects—enhanced hydrophilicity, favorable microstructural changes (reduced thickness), and modestly increased crosslinking—lead to higher water flux despite greater crosslink density. Charge effects from surface ammonium functionalities contribute to improved salt rejection via enhanced electrostatic exclusion of cations (Na+, Mg2+) and overall tighter transport pathways; divalent ions, especially MgCl2 and MgSO4, show the highest rejection. Anion interactions modulate flux, with more strongly attracted anions lowering flux via surface association and partial shielding. Under real seawater NF permeate conditions, the EDA-Deboc membrane sustains superior flux and TDS rejection compared to control and EDA-Boc variants, and exhibits robust antifouling behavior with high flux recovery after simple DI backwash. These results validate the hypothesis that in situ protection/deprotection chemistry enables precise active-layer tuning to advance both permeability and selectivity in RO membranes.

This study demonstrates a simple yet effective chemical strategy to tune polyamide RO membranes by incorporating Boc-protected ethylenediamine during IP and subsequently deprotecting in situ to generate surface ammonium groups. The optimized MPD-TMC-EDA-Deboc membrane delivered: (i) markedly enhanced hydrophilicity (WCA 55°), (ii) higher water flux (up to 37 LMH at 30 bar with DI water; 25 LMH with 2000 ppm NaCl at 20 bar), (iii) improved salt rejection (NaCl 98 ± 0.5%; divalent salts ≥98%), (iv) strong performance with real seawater NF permeate (97% TDS rejection, 23 LMH at 20 bar), and (v) excellent fouling reversibility (≈95% flux recovery after DI backwash). The approach avoids compatibility and defect issues associated with inorganic fillers by covalently integrating functionality within the polyamide network. These findings highlight in situ protection/deprotection as a versatile route to tailor active-layer chemistry for advanced desalination membranes. Future work could explore other protected amines, varied protection chemistries, and optimization of deprotection conditions to further refine charge density, crosslinking, and transport pathways.