ABSTRACT
Addressing the characterization of Natural Organic Matter (NOM) removal by functionalized membranes in water treatment, this study evaluates the effectiveness of two commercial ion-exchange membrane adsorbers: Sartobind® Q (with quaternary amines) and D (with tertiary amines). Using Suwannee River NOM (SRNOM) as a surrogate, Langmuir adsorption isotherms revealed maximum capacities (Qmax) of 2966 ± 153 mg C/m2 and 2888 ± 112 mg C/m2, respectively. Variations in flux from 50 to 500 LMH had a minimal impact on breakthrough times, proving low diffusion limitations. The macroporous (3-5 µm) functionalized cellulose-based membranes exhibited high permeabilities of 10,800 L/(h m2 bar). Q maintained positive zeta potential vs. pH, while D's zeta potential decreased above pH 7 due to amine deprotonation and turning negative above an isoelectric point of 9.1. Regeneration with 0.01 M NaOH achieved over 95% DOC regeneration for Sartobind® D, characterizing reversibility through a pH-swing. Cyclic adsorption showed that Q maintained its capacity with over 99% DOC regeneration, while D required acidic conditioning after the first regeneration cycle to mitigate capacity reduction and re-deprotonate the adsorber. These results have demonstrated the potential suitability of adsorber membranes, designed originally for biotechnological purposes, for the possible removal of disinfection byproduct precursors in drinking water treatment.
ABSTRACT
Ion adsorbing ultrafiltration membranes provide an interesting possibility to remove toxic ions from water. Furthermore, it is also possible to recover valuable elements. In this work, we demonstrate two easy strategies to modify polyacrylonitrile membranes with anion and cation adsorbing groups. The membranes were modified to have positively charged amine groups or negatively charged carboxyl groups. The success of the reactions was confirmed using IR spectroscopy and zeta-potential measurements. The membranes carrying negatively charged groups provided a negative zeta-potential and had an isoelectric point at pH 3.6, while the membranes carrying positively charged groups had a positive zeta-potential in the analyzed pH range. Since only the surface of the polymer was modified, the pore size and permeance of the membranes were not drastically affected. The membranes prepared by both modification strategies had a pure water permeance higher than 1000 L/(m2 h bar) and a water contact angle of 44.3 and 57.2°, respectively. Therefore, the membranes can be operated at low pressures with reasonable flux. Additionally, SEM images showed that the membranes were still open-pored. Adsorption tests using a positively and a negatively charged dye as well as a toxic cation and an anion were performed to analyze the adsorption behavior. Both membranes were able to adsorb the oppositely charged dyes as well as the copper and chromate ions. Therefore, these membranes are good candidates to purify water streams containing hazardous ions.
ABSTRACT
Negatively charged electrically conductive ultrafiltration (UF) membranes have been intensively investigated for fouling mitigation and rejection enhancement in recent years. This study reports the novel approach of applying positive charge (+2.5 V cell potential) to a conductive membrane to induce electrosorption of negatively charged substances onto the membrane. Subsequently, desorption of negatively charged substances is achieved by changing the potential periodically (e.g., after 30 min) to negative charge (-2.5 V cell potential). For this purpose, sputter deposition of ultra-thin gold layers (40 nm) is used to generate electrically conductive gold-polymer-gold flat sheet membranes by coating the active and the support layer of two commercial polymer UF membranes (polyethersulfone UP150, polyamide M5). When M5 membrane was charged positively during filtration (+2.5 V), Suwannee River NOM, Hohloh lake NOM, humic acid and Brilliant Blue ionic dye showed removal rates of 70 %, 75% and 93% and 99%, respectively. Whereas, when no potential was applied (0 V) removal rates were only 1 - 5 %. When a positive potential was applied to the active membrane layer and a negative potential was applied to the support layer (cell potential 2.5 V), a significant increase of flux with 25 L/(m² h) was observed due to the induction of electro-osmosis. Electrosorption was only observed for M5 membrane (ζ: +13 mV, pH 7) and not with UP150 membrane (ζ: -29 mV, pH 7). Due to a low current density of 1.1 A/m² at a flux of 100 L/(m² h), the additional energy consumption of electrosorption and desorption process was low with 0.03 kWh per m³ of permeate. This study delivered the proof of concept for the novel process of electrosorptive UF with energy consumption between microfiltration and ultrafiltration but NOM removal rates of nanofiltration membranes.
Subject(s)
Ultrafiltration , Water Purification , Humic Substances/analysis , Membranes, Artificial , WaterABSTRACT
Next to the pore size distribution, surface charge is considered to be one main factor in the separation performance of ultrafiltration (UF) membranes. By applying an external surface potential onto an electro-conductive UF membrane, electrostatic induced rejection was investigated. This study introduces in a first part a relatively simple but yet not reported technology of membrane modification with direct current sputter deposition of ultrathin (15 nm) highly conductive gold layers. In a second part, characterization of the gold-coated UF flat sheet membrane with a molecular weight cut-off (MWCO) of 150 kDa is presented. Membrane parameters as contact angle (hydrophobicity), pure water permeability, MWCO, scanning electron microscopy imaging, zeta potential, surface conductivity and cyclic voltammetry of the virgin and the modified membrane are compared. Due to the coating, a high surface conductivity of 107 S m-1 was realized. Permeability of the modified membrane decreased by 40% but MWCO and contact angle remained almost unchanged. In a third part, cross-flow filtration experiments with negative charged Suwannee River Natural Organic Matter (SRNOM) are conducted at different cathodic and anodic applied potentials, different pH values (pH 4, 7, 10) and ionic strengths (0, 1, 10 mmol L-1). SRNOM rejection of not externally charged membrane was 28% in cross-flow and 5% in dead-end mode. Externally negative charged membrane (-1.5 V vs. Ag/AgCl) reached rejection of 64% which was close to the performance of commercial UF membrane with MWCO of 5 kDa. High ionic strengths or low pH of feed reduced the effect of electrostatic rejection.