Fourth-generation flotation: the bubble is the machine
Every generation of flotation has been an argument about bubble size. Dissolved air settled at 20–100 µm. The fourth generation starts at 40 nm, and the rest of the design follows from that.
Four generations in one table
| Generation | Bubble size | Separation | Weak point |
|---|---|---|---|
| 1st · Dissolved air (DAF) | 20–100 µm | Pressurised dissolution, release nozzle | Low collision efficiency on fine particles and emulsified oil; 20–30% recycle ratio |
| 2nd · Shallow layer | 20–60 µm | Shallow tank, short retention | Smaller footprint, same bubble floor |
| 3rd · Compact swirl (GEM type) | 10–50 µm | Swirl contact, compact vessel | Better kinetics; still limited by gas dissolution |
| 4th · ARFU (SiC + swirl + staged) | 40 nm – 50 µm | Membrane gas generation, swirl cutting, staged injection, zero-velocity separation | Membrane is a consumable; feed solids need pre-screening |
What the four stages buy you
94% TSS removal
Staged injection extends gas–liquid–solid contact along the flow path. Measured at Cui Jiagou mine water; conventional DAF typically lands near 85% on the same duty.
0.18 kWh/m³ measured
Because operating pressure is 0.1–0.5 bar, not 4–6 bar. The high-pressure dissolution tank is the single biggest energy consumer in a classic DAF, and it is gone.
50% of the footprint
Zero-velocity separation needs less area for the same rise rate, and the absence of a saturation vessel removes a whole skid.
30–50% less sludge
A coherent float layer carries less water. That is a disposal cost line, not a cosmetic difference — haulage is often the second-largest operating cost after energy.
Where the numbers came from, and where they stop
Cui Jiagou mine water, 2×75 m³/h, accepted 2025: effluent 0.18–0.20 NTU against a 0.5 NTU reuse limit, TSS removal 94%, specific energy 0.18 kWh/m³ · Keshiketeng oily wastewater, 50 m³/h (ARFU-2025-099) · 500 t/h nanobubble flotation system, TSS removal 99.8% on a high-solids feed. The 94% figure is a mine-water result. On emulsified oil with high surfactant load, expect less, and require a bench test before you commit to a guarantee.
Retrofit: keep the basin, change the heart
Most plants with a flotation problem do not need a new basin. They need a new gas source. Three tiers, in increasing scope and cost:
| Tier | Scope | Downtime | Typical result |
|---|---|---|---|
| Tier 1 · Drop-in | Replace the release system with a nanobubble generator; keep tank, scrapers, and civil works | 3 days | Bubble size falls by two orders of magnitude; chemical dosing typically down 30–50% |
| Tier 2 · Contact zone | Tier 1 plus a staged injection manifold and a revised contact zone | 1–2 weeks | TSS removal into the 90%+ band on suitable feed |
| Tier 3 · New unit | Complete ARFU skid alongside or replacing the old basin | 4–6 weeks | Full 94% / 0.18 kWh/m³ performance, 50% footprint |
Downtime figures assume the existing basin can be isolated. On a single-train plant with no bypass, add the tie-in time — and tell us about it early, because it changes the recommendation more than the water chemistry does.
When a retrofit is the wrong answer
- The basin is hydraulically short-circuiting. Fix the baffles first; a better bubble will not fix a dead zone.
- Solids loading exceeds what any flotation unit can float — pre-settle or screen instead.
- The failure is downstream: oil in the effluent is emulsified below 10 µm and stabilised by surfactant. Flotation lifts free oil; dissolved oil needs chemistry or membrane.
Send a photo of the existing basin and the last water analysis
That is usually enough to say whether Tier 1 will do, or whether you are looking at a new unit.
Discuss a retrofit → The membrane inside it