Refineries Are Leaning on DAF to Crack Stubborn Oil Emulsions

Refinery effluent typically contains hundreds of mg/L of oil (and fine solids) that form stable emulsions not easily removed by gravity. Simple API (American Petroleum Institute) separators can remove free oil, but emulsified oil droplets—often less than 20 μm (micrometers)—resist settling. In practice, primary equipment such as screens and oil removal systems set the stage, but dissolved‑air flotation (DAF) is the workhorse for the stickiest fraction.

Here is why: flotation “attaches” tiny oil droplets to microbubbles, boosting their buoyancy. Studies have shown DAF is “more effective for the removal of emulsified‑oil droplets” than gravity separation because bubble attachment greatly increases rise speed (researchgate.net). That is why a modern DAF system is often specified as pretreatment ahead of polishing steps.

Refinery DAF operating envelope

Typical refinery DAF systems operate at 3–5 atm (atmospheres) saturation pressure, recycling roughly 10–20% of flow to generate microbubbles, with total hydraulic residence time (HRT, time the water spends in the tank) of about 10–20 minutes (cnwwtp.com). Design guides—citing Chinese standard HJ2007‑2010—recommend shallow tanks less than 2.5 m deep with long, narrow flow paths (length‑to‑width L:W ~2–3:1) and a surface overflow load around 4–6 m³/(m²·h) (cnwwtp.com). Surface overflow load is simply the flow rate per unit of tank surface area.

One guideline breaks the basin into a contact zone and a separation zone with ≤1 minute contact and 10–20 minutes separation, giving 4–6 m³/m²·h surface loading (cnwwtp.com). Free oil removal upstream—using equipment similar to dedicated oil skimmers—helps the DAF focus on the emulsified fraction.

Chemical pretreatment for emulsion break

Because emulsified oil droplets carry electrical charges and are colloidally stabilized, chemical pretreatment is usually required to break the emulsion and form floatable flocs. Coagulants (typically multivalent salts or pre‑hydrolyzed aluminum/iron products) work by charge neutralization, while polymeric flocculants bridge particles into larger aggregates. In practice, combinations are common. For example, Zhaoyang You et al. screened several coagulants for ship‑emulsion wastewater and found polyaluminum ferric chloride (PAFC) > poly‑aluminum chloride (PAC) > PSAFS > alum > Al₂(SO₄)₃ > poly‑ferric sulfate > FeCl₃ in demulsification capacity (pmc.ncbi.nlm.nih.gov).

They then dosed 300 mg/L PAFC + 300 mg/L PSAFS + 30 mg/L cationic polyacrylamide (PAM) at pH 6.6, achieving effluent oil less than 15 ppm despite 3000–5000 mg/L influent oil (pmc.ncbi.nlm.nih.gov) (pmc.ncbi.nlm.nih.gov). For aluminum‑based programs and hybrids, plant teams often draw from solutions like PAC/ACH coagulants and then add a dedicated flocculant to enhance bridging.

Jar tests on typical refinery wastewater echo that playbook. Albarazanje and Mohamed (2019) report DAF oil removal jumping from ~60% (no chemical) to ~90–94% with 25–50 mg/L PAM or alum alone, and reaching 98.9% (for 50 ppm influent) when using both (doaj.org). As one study put it, “PAM with added water was used to further reduce the oil content” after coagulation (pmc.ncbi.nlm.nih.gov).

Dosing optimization and control

Real‑world DAF performance is highly sensitive to chemical dose. Optimization studies consistently find that moderate doses (tens of mg/L) yield robust removal, while overdosing gives little extra benefit (and can even be counterproductive or wasteful). Tetteh (2018) optimized a pilot DAF for refinery oil and found >75% COD (chemical oxygen demand)/SOG (soluble oil and grease)/TSS (total suspended solids) removal at 50 mg/L coagulant (pH 5, 15 min, 10% recycle, 350 kPa) (openscholar.dut.ac.za). Response‑surface trials (RSM, a statistical method for process optimization) further showed ~85% oil‑SOG removal at only 30–45 mg/L coagulant (air ratio 8–12%, 300–425 kPa) (openscholar.dut.ac.za).

Notably, the continuous system required ~100–180 mg/L to hit similar removal, highlighting scale effects (openscholar.dut.ac.za). Foukalas & Galassi (2020) likewise report that DAF fed with 50 mg/L of aluminum‑ or ferric‑based coagulants (pH 5, 400 kPa) removed about 73–91% of TSS and 81–96% of turbidity, and 83–96% of COD (and similarly high oil/grease reduction) (mdpi.com). In all such studies, tuning the dose is done via jar testing or pilot RSM trials, since the optimum depends on influent quality.

Underdosing leaves residual oil and fines. Overdosing yields marginal gain at high chemical cost, and can create excess sludge or cause coagulant “restabilization” (charge reversal) (pmc.ncbi.nlm.nih.gov) (researchgate.net). In fact, You et al. note that FeCl₃ performed poorly at 500 mg/L (effluent oil ∼53 mg/L) and got worse as dose rose (pmc.ncbi.nlm.nih.gov). Many plants therefore rely on on‑line tests—turbidity monitors, streaming current meters, periodic zeta‑potential (particle surface charge) checks—to adjust alum/PAC and polymer doses in real time, often metered with an accurate dosing pump. The Iraqi study also provided equations linking removal to coagulant dose for plant planning (doaj.org).

DAF unit sizing and hydraulics

Beyond chemicals, DAF design itself drives efficiency. Key parameters include air saturation pressure, recycle ratio, and bubble size. All else equal, higher pressurization—e.g., up to 4–5 atm—dissolves more air and generates denser microbubbles when flashed. Typical coagulation/DAF tests show oil removal improving with higher saturator pressure and higher recycle percentage (doaj.org). Longer HRT (10–20 minutes) gives flocs time to rise.

Most DAF units have a contact/saturation zone and a separation zone: the former mixes coagulant with recycled stream (often with a 70° slanted baffle) (cnwwtp.com). Good design targets a flotation surface loading ≈0.5–1.0 m³/m²·h (4–6 m³/m²·h in Chinese units (cnwwtp.com)) so that bubble rise velocity (~0.5–2 cm/s) adequately carries flocs to the surface. As rule‑of‑thumb geometry, designers use less than 6 m width for horizontal DAF cells, 2–15 m length, 1–2 minutes contact time, and 10–20 minutes separation (cnwwtp.com). These principles underpin how a compact DAF achieves refinery‑class throughput.

Measured outcomes and efficiency

With optimized chemicals and design, refinery DAF can achieve very high oil‑water separation. Case studies and lab tests report >90% reduction of oil & grease and turbidity. Albarazanje et al. saw oil drop from 50 ppm to ≈0.5 ppm using alum+PAM (doaj.org). You et al. obtained effluent oil “stable below 15 ppm” from influent 3000–5000 mg/L after optimized PAFC+PAM dosing (pmc.ncbi.nlm.nih.gov).

Foukalas & Galassi’s data imply typical DAF will polish turbidity and COD by 80–95% given 50 mg/L coagulant (mdpi.com). These outcomes easily meet stringent discharge goals (e.g., many jurisdictions require less than 20–50 mg/L oil), while also producing relatively compact sludge volumes. Critically, each additional mg/L of chemical beyond the optimum yields diminishing returns: one would not, for example, double a 50 mg/L dose to try to hit 100% removal when 95% is achieved already. Instead, operators balance removal targets against chemical cost and sludge generation. That balance informs selection of coagulants and downstream aids.

Bottom line and standards

In summary, integrating coagulation/flocculation with DAF is essential for emulsified oil removal in refinery wastewater. Coagulant choice (PAC/PAFC or ferric) and a small dose of high‑molecular‑weight polymer can yield >85–99% oil/solids removal (doaj.org) (openscholar.dut.ac.za). Design guidelines (e.g., HJ2007‑2010) and jar‑test data should be used to size the flotation tank (typical load ~4–6 m³/m²·h, retention ~10–20 min (cnwwtp.com)) and to set recycle ratio (often 10–20%) and saturator pressure (~3–5 atm). Systematic dose optimization (e.g., via bench RSM studies) is critical: small changes in coagulant or polymer dose can substantially change effluent quality. Ultimately, well‑designed DAF with optimized dosing consistently reduces oil and fine solids to low mg/L levels, enabling compliance. Where needed, utilities pair these units with ancillary wastewater equipment for skimming and sludge handling.

Sources: Authoritative studies and industry reports (e.g., Zouboulis and Avranas 2000; You et al. 2018; Albarazanje & Mohamed 2019; Tetteh 2018; Foukalas & Galassi 2020) and technical guidelines (HJ‑2007‑2010) were used for data on removal efficiencies, coagulant selection, and design parameters (researchgate.net) (doaj.org) (openscholar.dut.ac.za) (mdpi.com) (cnwwtp.com). These provide quantitative evidence (removal %, optimal doses, effluent concentrations) to guide DAF design and operation.