A team at Hankyong National University in South Korea has compared organic–inorganic composite fouling on ceramic and polymer separation membranes in real semiconductor wastewater and published a material-specific mechanism, the National Research Foundation of Korea announced on 9 September 2026 (DongA Science). The work, led by Professor Noh Ho-jung of the Department of Civil and Environmental Engineering, is intended to support more stable, cost-effective reuse of fab wastewater (DongA Science).
Key takeaways
- Real fab wastewater was filtered for 48 hours on alumina ceramic and PES polymer hollow-fibre membranes, and the foulants were compared head-to-head (DongA Science).
- On both membrane types, silica–protein complexes and metal ion–organic complexes were identified as the main foulants (DongA Science).
- Final permeate flux on the ceramic membranes was only about one-third of that on the polymer membranes, counter to the common assumption that ceramics foul less (DongA Science).
- A new material-specific pathway was identified, in which iron ions adsorb onto Al–OH groups on the ceramic surface and then accumulate humic substances on top (DongA Science).
- The peer-reviewed paper appeared online in the international journal Desalination on 19 August (doi 10.1016/j.desal.2026.120686) (DongA Science).
What happened
As semiconductor fabs get larger, the volume of process wastewater they generate rises, and the case for purifying and reusing it has strengthened, with ultrafiltration (UF) and reverse osmosis (RO) membranes the workhorse technologies (DongA Science). RO pores are quoted at 0.0001 μm, while UF removes bacteria, proteins and colloids through pores tens of thousands of times thinner than a human hair (DongA Science). When contaminants adhere to and clog these membranes, flux drops, cleaning frequency rises, and downstream stability suffers — particularly with composite fouling built from silica, metal ions and organic matter, which has been hard to treat because earlier work used synthetic feeds or single foulants rather than real fab effluent (DongA Science).
To address that gap, Professor Noh's group ran 48-hour parallel tests on real semiconductor wastewater using alumina ceramic membranes and polyethersulfone (PES) polymer hollow-fibre membranes (DongA Science). Surface analyses showed the same two foulant families dominated on both materials — silica–protein complexes and metal ion–organic complexes — but the ceramic membrane ended the run at roughly one-third the final permeate flux of the polymer membrane (DongA Science). The team attributes that gap to a newly identified, material-specific fouling pathway on ceramics: iron ions selectively adsorb to Al–OH functional groups on the membrane surface, after which humic substances accumulate on top, so surface chemistry — not just pore structure — drives composite fouling (DongA Science).
"We have provided a scientific basis for selecting appropriate membrane materials by considering the chemical interactions between semiconductor wastewater components and membrane surfaces," said Professor Noh Ho-jung, adding that the results can be used to "design purification processes that jointly consider the composition of iron, silica, and organic matter in wastewater" (DongA Science).
Specification read
The reported result is not a flow figure but a flux ratio and a fouling mechanism, so the buyer-relevant scale is the test duration and the membrane surface area needed to draw conclusions about ceramic versus polymer performance. The 48-hour bench run on real fab effluent maps to a typical pilot-scale acceptance test for an UF or RO pre-treatment module (general industry range, not from the sources). On a 48-hour run the ceramic membrane's final permeate flux was about one-third of the polymer membrane's, even though ceramics are often assumed to foul more slowly (DongA Science). For a fab reuse train, the work touches the membrane stage that sits ahead of RO — usually fed by physico-chemical clarification such as DAF or a lamella clarifier to strip metals, silica and organics, and protected downstream by RO with chemical-clean-in-place skids; the study's iron- and silica-driven composite fouling is exactly the failure mode that stage must be designed to control. Operators specifying an Mbr Membrane Systems package for fab-side reuse should treat ceramic selection as a chemistry decision, not just a durability one. The findings also align with the engineering questions raised in our earlier DAF or Clarifier for Semiconductor Wastewater in South Hadley, MA: 2026 Factory Guide, where front-end solids removal has to be tuned to the foulants that follow the membrane.
FAQ
What does the result change for a plant choosing between ceramic and polymer UF membranes on fab wastewater?
It does not give a single winner; it says the choice should be driven by feed chemistry, because ceramic membranes can foul more than polymers in some real fab streams and the loss in flux has a specific chemical cause (DongA Science).
Which foulants should operators monitor most closely on a fab UF/RO reuse train?
Silica, iron, and organic matter — particularly silica–protein complexes, metal ion–organic complexes, and humic substances that build up on iron that has already adsorbed to the membrane surface (DongA Science).
What is a reasonable pilot duration to compare membrane materials on real fab wastewater before procurement?
The published study used a 48-hour run on real fab effluent; pilot work in this class is typically run for at least several days to expose flux decline (general industry range, not from the sources).
Where can I read the original peer-reviewed paper?
The work was published online on 19 August in the international journal Desalination, doi 10.1016/j.desal.2026.120686, as reported by DongA Science.
