Steel’s Oily Secret: DAF vs. Ultrafiltration in the Race to Clean Mill Water

From rolling draft to quench baths, steel plants generate lubricated, greased wastewaters laced with emulsified oils—droplets finer than 20 µm that resist simple gravity separation. Conventional gravity-separation fails on these fine emulsions, which is why plants lean on physico‑chemical flocculation/flotation or advanced membranes (MDPI) (LinkedIn).

One of the most installed units is DAF (dissolved‑air flotation, a micro‑bubble flotation process) that lifts flocculated droplets into a scum layer. On the other side is UF (ultrafiltration, a pressure‑driven membrane step with typical membrane molecular weight cut‑off, MWCO, of ~10–100 kDa), which sieves out droplets and suspended solids outright.

How they stack up—and how chemistry tips the balance—matters for footprint, chemicals, and the oil-and-grease (O&G) numbers mills must hit.

Oily wastewater characteristics and limits

Steel mills’ emulsified oils (<20 µm droplets) from rolling draft and quench baths drive the need for treatments beyond gravity (MDPI) (LinkedIn). Ahead of core treatment, mills often deploy primary physical steps—screens and free‑oil skimming—where equipment such as primary separation systems, an automatic screen, or targeted free‑oil removal can set up downstream performance.

DAF performance with and without chemistry

DAF injects micro‑bubbles that attach to flocculated droplets and carry them to the surface. In practice, DAF alone typically removes only ~50–70% of emulsified oil; an Iraqi refinery study found 60% oil removal with no coagulant (ResearchGate).

The fix is chemistry. With 25–30 mg/L cationic polyacrylamide (PAM) in DAF, oil removal reached ~94% (from 50 ppm influent), versus 89% with alum alone—with a blend of alum + PAM achieving ~99% removal (ResearchGate). Typical DAF effluent after such treatment may still contain tens of mg/L of oil/grease; pilot data report UF permeates <100 mg/L and TSS (total suspended solids) very low (ResearchGate).

DAF requires a sizeable footprint (large tanks) and continuous sludge (scum) removal. Energy costs are modest, but chemical costs (coagulants) can be high (ResearchGate) (LinkedIn). Where appropriate, integrated units like a compact DAF system help concentrate scum efficiently for removal.

Ultrafiltration membranes under load

UF membranes—typically polymeric hollow‑fiber or flat‑sheet with ~10–100 kDa MWCO—retain almost all oil droplets and suspended solids. Pilot/full‑scale trials at refineries and factories consistently show very high oil rejection: Ostlund et al. reported permeate O&G <100 mg/L under continuous operation (ResearchGate).

In one study with two spiral UF modules (MWCO 2–35 kDa), the finer membrane delivered ~90% COD (chemical oxygen demand) and ~99.7% hydrocarbon (HC) removal at a permeate flux ~20 L/m²·hr, 400 kPa, 35°C (ResearchGate). A larger refinery UF project found O&G rejections ≈97–98%, with effluent typically <50 mg/L oil/grease (ResearchGate) (ResearchGate).

UF permeate easily meets <100 mg/L oil targets and offers a compact footprint—often chosen for “polishing” after primary treatment (MDPI). Plants frequently specify skid systems such as ultrafiltration units with proven UF/RO membranes to fit tight spaces.

Tradeoffs are real: UF is sensitive to oil fouling—pores clog and surfaces foul as oils adhere—requiring regular cleaning and chemical clean‑in‑place (LinkedIn) (ResearchGate). Fouling can be partially mitigated by pretreatment (e.g., skimmers/DAF to remove free oil) and membrane surface modification (hydrophilic coatings). Despite cleaning, UF’s low operating pressure (~0.2 bar) makes it energy‑efficient at ~0.1–0.2 kWh/m³ (MDPI). Plants often pair cleaning programs with compatible agents, for example through dedicated membrane cleaners.

Head‑to‑head: efficiency, footprint, chemicals

Removal efficiency: DAF with optimized coagulant dosing can approach UF levels—~94–99% oil removal when dosed with PAM and/or alum (ResearchGate)—while UF tests consistently show ~97–99% oil rejection (ResearchGate) (ResearchGate).

In absolute terms, UF permeate concentrations are often <50 mg/L O&G (ResearchGate), whereas DAF effluent typically needs further treatment to reach such levels. Footprint and cost: UF systems have much smaller hydraulic footprint (though higher capital cost per volume), while DAF units are bulky and require large sludge handling. Chemicals: DAF needs coagulant/flocculant feed to achieve high removal; UF needs none except periodic cleaning—yet UF requires upstream oil removal (typically via skimming/DAF or coalescers) to prevent severe fouling (LinkedIn). In contrast, DAF demands larger volumes for treatment and more chemical use, though DAF is robust for variable loads (ResearchGate) (LinkedIn).

Demulsifiers and flocculants: breaking stability

Stable oil‑water emulsions must be broken (demulsified) and droplets aggregated (flocculated) before removal. Demulsifiers—typically oil‑soluble surfactants or polymers—penetrate and destabilize the surfactant film on droplets, reducing interfacial tension and neutralizing surface charge; once demulsified, flocculants (often cationic polyacrylamides or coagulants like Fe/Al salts) neutralize remaining charge and bridge droplets into flocs, driving coalescence and flotation (MDPI).

Example: adding 100 mg/L of an oil‑soluble demulsifier cut the oil–water interfacial tension from ~4.43 to 1.33 mN/m and reduced the droplet zeta potential (a measure of droplet surface charge) by ~60% (from –67.1 to –30.6 mV), greatly destabilizing the emulsion (MDPI) (MDPI). In practice, compounds used include cationic organic polymers (e.g., high‑molecular‑weight polyacrylamides, poly(diallyldimethylammonium chloride)), sometimes grafted onto magnetic particles or supports for easy removal, and inorganics like polyaluminum chloride (PAC) or ferric chloride. For DAF specifically, adding just 25 ppm PAM raised oil‑removal to ~94% (ResearchGate).

Novel demulsifier‑based flocculants are pushing doses lower: a magnetic chitosan‑polymer composite (FS‑MC) at only 2–3 mg/L removed ~93–95% of emulsified oil, and a polymer blend (2–4 mL/L) achieved ~95.8% oil removal from steel rolling wastewater (ResearchGate) (ResearchGate). These flocculants combine hydrophobic and cationic groups to maximize bridging and charge neutralization.

In real systems, demulsification + flocculation upstream of DAF (or pretreatment to UF) achieves very low oil concentrations. Bench experiments report oil “clarity” (TCOD or turbidity) drop by >90% after optimized demulsifier+flocculant dosing, with dosages typically in the 1–100 mg/L range depending on oil load; jar tests tailor the pick to make droplets big enough to float effectively (ResearchGate) (ResearchGate).

Operationally, accurate chemical feed with a dosing pump and targeted selections from demulsifiers, coagulants, and flocculants underpin consistent separation.

What the studies say—one view

Modern reviews and studies confirm the main trends: UF membranes routinely exceed 97% oil rejection (ResearchGate), DAF with coagulant sits near 95–99% (ResearchGate), and advanced flocculants/demulsifiers can achieve >90% removals at very low dose (ResearchGate) (MDPI). In contrast with compact UF, DAF demands larger volumes and more chemical use, though it is robust for variable loads (ResearchGate) (LinkedIn).

Study references

• Albarazanje, M. G. et al. Study the Performance of Dissolved Air Flotation as Industrial Wastewater Treatment Method, Eng. Techn. J. 37‑C(3), 333–343 (2019) (ResearchGate).
• Reed, B. et al. Oily Wastewater Treatment by Ultrafiltration: Pilot‑Scale Results and Full‑Scale Design, Pract. Period. Hazard. Toxic Radioact. Waste Mgmt. 2(3), 100–110 (1998) (ResearchGate) (ResearchGate).
• Huang, B. et al. Demulsification–Flocculation Mechanism of Oil–Water Emulsion, Polymers 11(3), 395 (2019) (MDPI).
• Chen, J. H. et al. Cov. Bonded Magnetic Flocculant for Emulsified Oil Removal, J. Hazard. Mater. 410, 124367 (2021) (ResearchGate).
• Abuhasel, K. et al. Oily Wastewater Treatment: Conventional and Modern Methods, Water 13(7), 980 (2021) (Review article) (MDPI).