The researchers, Hadi Rouhi and Steven T. Weinman of the Department of Chemical and Biological Engineering, started with commercial polyvinylidene difluoride, or PVDF, ultrafiltration membranes. PVDF is a favorite in water treatment because it shrugs off oxidants, halogens, and most acids and solvents, but its hydrophobic surface gives it little natural affinity for charged pollutants. The team functionalized the membranes with three diamines: 1,3-propanediamine carrying two primary amines, N-methyl-1,3-propanediamine carrying a primary and a secondary amine, and N,N-dimethyl-1,3-propanediamine carrying a primary and a tertiary amine. A fourth variant was created by quaternizing the tertiary amine with iodomethane, producing a permanently charged quaternary ammonium group. The membranes are labeled P, S, T, and Q according to their amine type.

The chemistry behind the modification is itself interesting. Rather than directly displacing the carbon-fluorine bonds of PVDF, the diamines act as bases that drive dehydrofluorination, stripping hydrogen fluoride from the polymer backbone and creating conjugated polyene segments. The amines then attach to some of these newly formed carbon-carbon double bonds through a Michael addition. The visual evidence is striking: the initially white membranes turn brown, darkening progressively from P to Q as more dehydrofluorination occurs. A control experiment with sodium hydroxide, which produces a similar pH but no amines, confirmed that the polyene segments alone do not capture DBSA; it is the amine functionality that does the work.

Characterization backed this up. Attenuated total reflectance Fourier transform infrared spectroscopy revealed new absorption bands at 1532 and 1630 wavenumbers, assigned to N-H bending and C-N stretching, in every diamine-modified membrane. X-ray photoelectron spectroscopy detected a distinct nitrogen 1s peak at 398 electron volts, and the quaternized membrane showed iodine signatures from the methyl iodide reagent. Scanning electron microscopy showed that surface morphology was essentially unchanged, consistent with a monolayer attachment rather than a bulky multilayer coating. Water contact angles shifted only slightly, from 93.2 degrees for unmodified PVDF to 83.0 degrees for the P membrane, again reflecting the thin functional layer.

Performance testing told the central story. Challenged with a demanding DBSA solution of roughly 600 milligrams per liter, far above the 15 to 400 milligrams per liter range typical of adsorption studies, the P, S, and T membranes removed 93.1, 91.4, and 95.3 percent of the surfactant within three hours. The quaternary Q membrane lagged badly, capturing only 37.3 percent in three hours and needing a full day to reach 72.1 percent. Adsorption capacities climbed from a meager 3.9 milligrams per gram for unmodified PVDF to as much as 22.2 milligrams per gram for the modified membranes. The tertiary amine membrane emerged as the fastest and strongest performer, a result the authors attribute to its higher effective positive charge density and the absence of intramolecular hydrogen bonding that would otherwise tie up active sites.

Kinetic analysis added a subtle twist. The P, S, and Q membranes followed pseudo-second-order kinetics, with coefficients of determination of 0.99, while the T membrane was uniquely described by pseudo-first-order kinetics. The team used the corrected Akaike Information Criterion on nonlinear residuals to make the call statistically, a rigorous touch that guards against artifacts of linearization. They interpret the two-stage adsorption profile of the P and S membranes as rapid electrostatic uptake followed by slower, orientation-dependent hydrogen bonding, whereas the T membrane interacts with DBSA through relatively unhindered ion-pairing. The Q membrane showed a sigmoidal profile, suggesting that its bulky trimethylammonium head group and ordered hydration shell must be partially displaced before effective binding can occur.

Isotherm studies pointed to heterogeneous surface interactions. The Temkin model provided the best fit for the P, S, and T membranes, with the adsorption heat constant sitting near 400 to 500 joules per mole for all variants, a hallmark of physisorption. The Sips model, which blends Langmuir and Freundlich behavior, gave the highest capacity and best fit for the T membrane at 21.3 milligrams per gram. Solution pH mattered enormously: removal efficiencies peaked under acidic conditions, where the amine groups are protonated and electrostatic attraction is strongest, and declined as pH rose toward 11, though appreciable removal persisted even there, hinting at hydrophobic and dipole contributions. Background salts of 200 milligrams per liter of sodium chloride or sodium sulfate trimmed removal by roughly 7 to 15 percent, with the T membrane the most ion-sensitive.

Thermodynamics sealed the mechanistic picture. Adsorption was exothermic on every modified membrane, with enthalpy changes of negative 20.0, negative 16.6, negative 27.5, and negative 2.9 kilojoules per mole for the P, S, T, and Q membranes respectively, and isosteric heats ranging from 2.1 to 25.4 kilojoules per mole. All values fall comfortably below the 40 kilojoules per mole threshold conventionally associated with physisorption, confirming that electrostatic attraction and hydrogen bonding, not chemical bond formation, govern the capture. Adsorption became less favorable at elevated temperatures, and the positive entropy changes reflect the release of structured water molecules from the DBSA sulfonate head group and the charged membrane surface as ion-pairs form.

Regeneration results were promising for real-world use. After five adsorption-desorption cycles, using an n-hexane and 1-nonanol mixture to strip the surfactant from the surface, the P, S, T, and Q membranes retained removal efficiencies of 69.9, 68.6, 70.6, and 72.0 percent respectively, down from first-cycle values of 86.6, 88.8, 87.4, and 72.0 percent. Notably, the Q membrane showed almost no decline across cycles, suggesting that its weaker but more reversible interactions resist site fouling. Dynamic flow testing, however, exposed a practical limitation: under continuous-flow conditions with limited residence time, the monolayer functionalization saturates quickly, and the authors suggest that future work should increase surface amine density, perhaps through a thin polymer interlayer, to bridge the gap between static and dynamic performance.

The broader significance is that amine structure is not a cosmetic detail; it is a design lever. Primary and secondary amines bind strongly but can be harder to regenerate, tertiary amines offer the best balance of speed, capacity, and reversibility, and quaternary ammonium groups trade strength for consistency across pH. As surfactant pollution continues to strain sewer networks and treatment plants worldwide, and as roughly one-third of the global urban population already lacks safe drinking water, membranes that can be tuned at the molecular level to capture specific anionic contaminants represent a genuinely exciting direction. This study provides the systematic evidence base for making that tuning rational rather than trial-and-error.

Subject of Research: Primary, secondary, tertiary, and quaternary Alkyl amine-functionalized PVDF membranes for DBSA removal

Article Title: Primary, secondary, tertiary, and quaternary Alkyl amine-functionalized PVDF membranes for DBSA removal

Article References: Rouhi, H., & Weinman, S. T. (2026). Primary, secondary, tertiary, and quaternary Alkyl amine-functionalized PVDF membranes for DBSA removal. Frontiers in Membrane Science and Technology, 5, Article 1892716. https://doi.org/10.3389/frmst.2026.1892716

Image Credits: AI Generated

DOI: 10.3389/frmst.2026.1892716

Keywords: Primary, secondary, tertiary, quaternary, Alkyl, amine-functionalized, PVDF, membranes, DBSA, removal, scientific research