Improved disinfection byproduct removal using a polysulfone membrane loaded with powdered activated carbon

Abstract

Powdered activated carbon was immobilized by casting it in a polysulfone polymer membrane, which was then tested for disinfection byproduct (chloroform) and bacteria (Escherichia coli) removal. The membrane prepared using 90% T20 carbon and 10% polysulfone (M20-90) provided a filtration capacity of 2783 L m À2 , adsorption capacity of 2.85 mg g À1 , and 95% chloroform removal in a 10 s empty bed contact time. Flaws and cracks on the membrane surface caused by the carbon particles appeared to reduce chloroform and E. coli removal. To overcome this challenge, up to six layers of the M20-90 membrane were overlapped, which improved chloroform filtration capacity by 94.6%, to 5416 L m À2 , and increased the adsorption capacity by 93.3%, to 5.51 mg g À1 . E. coli removal also increased from 2.5 logs reduction using a single membrane layer to 6.3 logs using six layers under 10 psi feed pressure. The filtration flux declined from 6.94 m 3 m À2 day À1 psi À1 for a single layer (0.45 mm thick) to 1.26 m 3 m À2 day À1 psi À1 for the six-layer membrane system (2.7 mm thick). This work demonstrated the feasibility of using powdered activated carbon immobilized on a membrane to improve chloroform adsorption and filtration capacity while simultaneously removing microbes. Practitioner Points• Powdered activated carbon was immobilized on a membrane to improve chloroform adsorption and filtration capacity while simultaneously removing microbes.• Membranes made with the smaller carbon particles (T20) delivered better chloroform adsorption performance.• Use of multiple layers of the membrane further improved chloroform and Escherichia coli removal.

The impact of PAC‐loaded polymer membrane thickness on chloroform removal and comparison of solvent and thermal membrane regeneration methods

“…As described by Hou and Mayer (2023), the membrane was produced by first attaching a 20 cm × 30 cm non‐woven fabric (NWF) substrate (AWA Paper, G80, Japan) to flat glass. Approximately 10 mL of solution was poured over the NWF, and a casting blade (MXBAOHENG, 500/750/1000/1500 μm, China) was used to form a thin film on the NWF.…”

“…The film was soaked in deionized (DI) water at 15°C for 24 h to allow the DMF to diffuse from the film into the water, forming a PAC membrane. The membrane alone, without PAC loading, provided less than 10% chloroform removal (Hou & Mayer, 2023), such that any further improvements in chloroform removal were attributed to the PAC loading.…”

“…There was no significant difference in the flux using M5, M10, and M15 (Tukey, p > 0.318), which ranged from 6.45 to 7.35 m 3 m À2 d À1 psi À1 (Figure 1c). Given that the flux through the membrane was not impacted by the membrane thickness, the majority of the pressure drop (or flow resistance) likely occurred at the membrane surface, which was composed of a dense PS layer, as observed by Hou and Mayer (2023) in SEM cross-sections of the PAC membranes. To test this hypothesis, we elected to use 1000-grit sandpaper (T417 Norton Saint-Gobain, USA) to polish the membrane surface to remove the top PS layer, targeting improved flux through the PAC membrane.…”

“…We previously developed a polysulfone membrane loaded with PAC and showed that it offered effective adsorption of chloroform (a representative THM) (Hou & Mayer, 2023). This PAC-loaded membrane is suitable for household water purification because of its rapid adsorption rate and higher adsorption capacity compared with household GAC block filters.…”

“…The PAC-loaded membrane surface had cracks that led to uneven flow of feed water, which in turn reduced adsorption capacity (Hou & Mayer, 2023). This issue was overcome by overlapping several layers of PAC membranes, which increased the total filter thickness and evened out the water distribution, but decreased flux, which may limit household applications.…”