One name, two processes: DAF as drinking-water clarifier and DAF as sludge thickener — the microbubble physics, the air-to-solids ratio that governs both, and why a design figure from one is meaningless in the other.
Every operator knows the failure: gas bubbles form inside the clarifier blanket, attach to floc, and float the solids you already captured. DAF reproduces that exact mechanism deliberately — bubbles supplied on purpose, and a skimmer waiting where the failure used to be.
Every operator knows rising sludge: denitrification generates nitrogen gas inside the clarifier blanket, the gas attaches to floc, and solids that settled properly float back to the surface. It is a defect. It gets written up and corrected. The New Mexico operator manual describes dissolved air flotation as relying on “a process similar to what happens to sludge during denitrification” — which is exactly that defect, reproduced deliberately, with the floating layer skimmed off as product.
That framing explains everything else about DAF, including the strangest fact about it: the same three letters describe two processes that barely overlap. In drinking water, DAF is clarification and the water is the product. In wastewater, DAF is thickening and the float is the product. Their loading rates differ by an order of magnitude, their recycle rates differ by a factor of twenty, and the solids range where one stops working sits far below where the other starts.
Gravity separation asks particles to be denser than water, and most of the time they oblige. Grit settles. Primary solids settle. Well-formed chemical floc settles. The entire architecture of conventional treatment — grit chambers, primary clarifiers, secondary clarifiers, gravity thickeners — rests on that assumption.
Some particles do not oblige. Algae are close to neutrally buoyant, and some species regulate their own buoyancy. Waste activated sludge is light and fluffy. Color-causing organic matter forms floc with almost no mass. Oil and grease are lighter than water to begin with. For these, waiting longer does not help — the settling velocity is not small, it is effectively zero or negative.
DAF does not make stubborn particles settle faster. It changes the sign of the problem. Attach enough air to a floc particle and the combined density of floc-plus-bubbles drops below that of water. The aggregate rises — faster and more reliably than it would ever have settled.
DAF is good at exactly one thing: anything that will not settle. Light particles, biological solids, cold water, low turbidity, oils. And it is bad at the mirror image: anything that settles well already. Pushing dense solids or high-turbidity raw water through a DAF means paying to dissolve air in order to lift material that would have fallen on its own.
A design figure quoted from one DAF is meaningless in the other. This is the single most common source of confusion in the literature.
| Water treatment: clarification | Wastewater: sludge thickening | |
|---|---|---|
| What it is | A clarification step replacing or supplementing sedimentation, ahead of filtration. | A solids-handling step concentrating sludge before digestion or dewatering. |
| The product | The water. The subnatant goes on to filters; the float is waste. | The float. The thickened sludge is the product; the subnatant returns to the plant. |
| Feed solids | Low. Not suited above about 100 NTU. | High. Minimum influent around 5,000 mg/L. |
| Hydraulic loading | 4–6 gpm/ft² conventional; 12–20 gpm/ft² high-rate. | 0.5–1.5 gpm/ft² (to 2.0 with polymer). |
| Recycle | Single digits to low tens of percent. | 100–200% of feed. |
| Success looks like | Low turbidity to the filters, long filter runs. | Float at 3–5% solids, subnatant under 100 mg/L TSS. |
Clarification DAF becomes uneconomical above roughly 100 NTU. Thickening DAF needs a minimum feed of about 5,000 mg/L. Under any reasonable correlation between turbidity and suspended solids, those figures sit more than an order of magnitude apart — a wide band where neither configuration is the right tool. That is not a gap in the technology; it is the signature of two processes developed separately, for different problems, that happen to share a working principle and a name. When you read a DAF design figure, establish which process it came from before you use it. A recycle rate of 150% is correct for a thickener and absurd for a clarifier.
The mass of gas dissolved in a liquid is proportional to the pressure of that gas above it. Raise the pressure and water holds more air. Drop the pressure and the water cannot keep what it was holding, so the excess comes out of solution as bubbles — the manual’s analogy is opening a can of soda, and it is the right one.
Water at 20°C dissolves roughly 18.7 mL of air per liter at atmospheric pressure, falling to about 15.7 mL/L at 30°C. A DAF saturator runs at 4–6 bar gauge (roughly 5–7 atmospheres absolute) and achieves 50–90% of full saturation. The air actually released on depressurization is the difference between what the stream held under pressure and what it can hold at atmosphere:
| Saturator pressure | Saturation fraction | Air released per liter of recycle |
|---|---|---|
| 5 atm absolute | 0.7 | 46.8 mL |
| 6 atm absolute | 0.7 | 59.8 mL |
| 7 atm absolute | 0.8 | 86.0 mL |
Released air forms microbubbles of roughly 30–100 µm — and that size is the design target, not an accident. Two independent reasons drive it. First, interfacial area: bubbles do their work at their surface, and surface area per unit volume scales as 6/d. A 50 µm bubble offers 40 times the interfacial area of a 2 mm coarse bubble carrying the same air — a single 2 mm bubble contains the same air as 64,000 bubbles at 50 µm. One arrangement can find and attach to floc; the other cannot. Second, rise velocity:
| Bubble diameter | Rise velocity | Behavior |
|---|---|---|
| 30 µm | 0.49 mm/s (1.8 m/hr) | Drifts upward. Long contact time with floc. |
| 50 µm | 1.36 mm/s (4.9 m/hr) | The typical working size. |
| 100 µm | 5.43 mm/s (19.5 m/hr) | Fast enough to begin shearing weak floc. |
| 2 mm (coarse) | Turbulent — meters per second | Tears through the water. Mixes rather than floats. |
Coarse-bubble aeration and dissolved air flotation both put air into water. One transfers oxygen through turbulence; the other attaches a buoyant particle to a fragile floc without breaking it. The bubble-size difference is not a detail — it is the difference between the two processes. Dispersed air flotation (mechanically sheared or diffused air) is a different technology for robust, hydrophobic targets like mineral processing; its bubbles are too large for hydrophilic chemical floc, and the turbulence destroys the floc you spent coagulant building.
Making microbubbles is a solved mechanical problem. Getting them to attach to the particles you want removed is a chemistry problem — and it is where DAF units fail. A bubble drifting past a stable, negatively charged, fully hydrated particle does not attach to it. Both bubble and particle carry negative surface charge and repel each other. Coagulation to near-zero charge removes that barrier, and it is the single most important chemical condition in the process. What flotation needs from coagulation is charge neutralization; what sedimentation needs is mass. Those are different objectives, and they produce opposite floc:
| Sedimentation wants | DAF wants | |
|---|---|---|
| Floc size | Large — settling velocity scales with diameter squared. | Small. Large floc is heavy, hard to lift, and easily sheared. |
| Floc strength | Moderate — it only has to survive transfer to the basin. | High. It must survive contact with the bubble stream intact. |
| Flocculation time | Long — typically 20–30 minutes to grow floc. | Short. Enough to destabilize and aggregate, not to grow. |
| Coagulant dose | Higher, to build settleable mass. | Lower. Heavy floc is not the goal. |
On a settling basin, a generous coagulant dose mostly costs money — bigger, heavier floc still settles. On a DAF the same instinct works against the process: heavier floc needs more air to lift, which demands more recycle and saturator energy, and past a point the floc simply does not float. It accumulates in the bottom of a tank that often has minimal provision for settled solids. A plant that converts from sedimentation to flotation and keeps its old jar-testing recipe will underperform for reasons that look mechanical and are not. Jar test for the minimum effective dose, not the comfortable one.
A/S is the master variable in DAF — the one parameter that connects the chemistry to the machinery. It is simply the mass of air released divided by the mass of solids fed:
Sa — air solubility (mL/L, falls as water warms) · f — saturation fraction achieved (0.5–0.9) · P — saturator pressure (atm absolute) · R — recycle flow · Sa,inf — influent suspended solids · Q — feed flow. The 1.3 converts air volume to mass. Everything in the numerator is equipment and operation; the denominator is what arrived at the plant. A/S is not something you set — it is something that happens, and the operator’s job is to move the numerator to track a denominator that changes without asking.
| Application | Typical A/S | Why |
|---|---|---|
| Oil, grease, dense hydrophobic solids | 0.005–0.02 | Already near-buoyant and attach readily. Very little air needed. |
| Light chemical floc | 0.03–0.05 | Low mass but poor natural attachment. Needs generous air. |
| Biological sludge | 0.02–0.06 | The general industrial range. |
| Sludge thickening (NM manual) | 0.01–0.1 | Wider band — unchanged by polymer addition. |
Part 2 left an unexplained fact on the table: clarification DAF runs single-digit recycle while the thickener spec says 100–200%. It falls straight out of the equation, because the denominator differs by two orders of magnitude. Take a saturator at 6 atm absolute and 70% saturation — about 59.8 mL, or 77.8 mg, of air per liter of recycle — and solve for the recycle each application needs:
| Application | Feed solids | Target A/S | Air needed | Recycle required |
|---|---|---|---|---|
| Drinking water | 30 mg/L | 0.05 | 1.5 mg/L | 1.9% |
| Drinking water | 60 mg/L | 0.05 | 3.0 mg/L | 3.9% |
| Secondary effluent | 150 mg/L | 0.04 | 6.0 mg/L | 7.7% |
| WAS thickening | 5,000 mg/L | 0.03 | 150 mg/L | 193% |
| WAS thickening | 8,000 mg/L | 0.03 | 240 mg/L | 309% |
This is also the practical meaning of the rule that A/S is inversely proportional to influent solids. Double the feed strength and you halve the achieved ratio at constant recycle. A thickener receiving an unexpectedly strong WAS is not merely loaded — it is underaerated, and the float will thin before anything shows on a loading chart.
On the potable side, DAF competes with sedimentation for the clarification step between flocculation and filtration — and it wins in a specific, well-defined set of conditions:
| Raw water condition | Why settling struggles | Why DAF works |
|---|---|---|
| Algae and cyanobacteria | Near-neutral buoyancy; some species regulate it actively. | Cells attach readily to bubbles and are carried up. |
| Low turbidity | Too few particles to build settleable floc. | Does not need mass. Small floc plus air is sufficient. |
| Color and organic matter | Organic floc is voluminous and almost massless. | Low density is an advantage rather than an obstacle. |
| Cold water | Higher viscosity slows settling; basins lose capacity in winter. | Less affected — and cold water dissolves more air. |
| Variable source quality | Basins respond slowly to changing raw water. | Short residence time; the unit responds within minutes. |
Sedimentation and flotation respond to temperature in opposite directions. Falling temperature raises viscosity and slows settling, so a basin loses capacity in winter — exactly when many surface waters are hardest to treat. The same falling temperature increases air solubility (18.7 mL/L at 20°C against 15.7 at 30°C), so the saturator dissolves more air per unit of energy. The two processes sit on opposite sides of the thermometer.
Conventional DAF clarifiers load at 4–6 gpm/ft²; proprietary high-rate designs reach 12–20 gpm/ft² — three to four times the throughput per square foot, bought at the cost of operating margin. A unit at 16 gpm/ft² has far less tolerance for a coagulation upset than one at 5. And the technology has a hard ceiling: above roughly 100 NTU, DAF becomes energy-intensive and uneconomical — past it, you are paying compressor energy to lift material gravity would remove for free.
On the wastewater side, the governing comparison is not against sedimentation but against the gravity thickener — and the two are not competitors. They are a matched pair, each taking the stream the other cannot handle:
| Gravity thickener | DAF thickener | |
|---|---|---|
| Primary sludge | Thickens best here. Dense, settles readily. | Less adaptable to flotation — and deposits grit that the unit is poorly equipped to remove. |
| Secondary sludge / WAS | Difficult — low solids and denitrification float the blanket. | The application. Light, fluffy solids that will not settle are exactly what flotation handles. |
| Sludge age | Older sludge gasifies and floats — a defect. | “Young” sludge thickens better than “old.” |
| Output | 2–4%, up to 6% with polymer. | 2–4%, 3–5% with polymer. |
Notice the two achieve nearly the same output concentration. The choice between them is not about how thick the product is — it is entirely about which sludge you are feeding. For where thickening sits in the solids train, see the thickening and dewatering guide.
| Operating parameter | Without polymer | With polymer |
|---|---|---|
| Solids loading, lb/hr/ft² | 0.4 – 1.0 | 1.0 – 2.0 |
| Hydraulic loading, gpm/ft² | 0.5 – 1.5 | 0.5 – 2.0 |
| Recycle, % | 100 – 200 | 100 – 200 |
| Air:solids ratio | 0.01 – 0.1 | 0.01 – 0.1 |
| Minimum influent solids, mg/L | 5,000 | 5,000 |
| Thickened sludge concentration, % | 2 – 4 | 3 – 5 |
Read across the rows and polymer’s effect is precise: solids loading capacity roughly doubles, hydraulic loading barely moves, recycle does not change, and air-to-solids does not change at all. Polymer is not buying air efficiency or hydraulic capacity — it buys solids capacity per square foot and about one extra point of float concentration. On a solids-limited unit, polymer roughly doubles what it can take. On a hydraulically limited unit, it will not help, and the dosing cost is wasted. Knowing which limit you are against is the difference between polymer being the answer and polymer being an expense.
Four levers, and none of them free:
| Lever | Raises | Costs |
|---|---|---|
| Recycle rate | Air delivered, therefore A/S. The primary operating control. | Pumping energy, and hydraulic load on the separation zone. |
| Saturator pressure | Air dissolved per unit of recycle. | Compressor energy, which rises faster than the air gained. |
| Coagulant / polymer | Attachment efficiency and solids capacity. | Chemical cost — and, if overdosed, floc too heavy to float. |
| Scraper speed | Removal rate; sets float blanket thickness. | Thinner float if too fast; carryover if too slow. |
The manual’s warning about the first two is worth repeating: the saturation ratio and the recycle rate depend on each other. Raising recycle against a fixed saturator can drop the saturation fraction achieved, so the air delivered rises less than the recycle does — and past a point, not at all.
| Symptom | Likely causes | Where to look first |
|---|---|---|
| Thin float, poor concentration | Excessive scraper speed; insufficient A/S; feed solids above design. | Slow the scraper first — it is free. Then check feed solids against the A/S actually achieved. |
| Solids in the subnatant | Blanket too deep and rolling under; carryover; poor attachment. | Blanket depth against the 6–8 inch target, then coagulation. |
| Solids settling to the bottom | Floc too heavy to float — often coagulant overdose; grit; A/S too low. | Jar test for the minimum effective dose, not the comfortable one. |
| Performance fell, nothing changed | Feed solids rose, so achieved A/S fell; or water warmed, so less air dissolves. | Feed concentration and temperature. Both move A/S without touching a control. |
| Large bubbles, boiling surface | Nozzle wear or fouling; saturator carrying undissolved air. | Release nozzles and saturator level control. |
| Float will not form at all | Charge not neutralized; floc sheared between flocculator and contact zone. | Jar testing, then the hydraulics of the transfer. |
When a DAF degrades without an operational change, calculate the A/S you are actually achieving rather than the one the unit was designed for. Two things move it without anyone touching a control: feed solids rising (the denominator grows) and water warming (the numerator shrinks). Both are seasonal, both are gradual, and neither trips an alarm. An operator who recomputes A/S monthly against measured feed solids and temperature will see the drift months before the float does.
| Figure | Value |
|---|---|
| Air solubility, 20°C, 1 atm | ~18.7 mL/L (15.7 mL/L at 30°C) |
| Microbubble diameter | 30–100 µm |
| Rise velocity, 50 µm bubble | ~1.4 mm/s (4.9 m/hr) |
| Saturator pressure | 4–6 bar gauge (~5–7 atm absolute) |
| Saturation fraction achieved | 0.5–0.9 |
| A/S, oil and grease | 0.005–0.02 |
| A/S, light chemical floc | 0.03–0.05 |
| A/S, biological sludge | 0.02–0.06 |
| Clarification loading, conventional | 4–6 gpm/ft² |
| Clarification loading, high-rate | 12–20 gpm/ft² |
| Clarification turbidity ceiling | ~100 NTU |
| Thickening hydraulic loading | 0.5–1.5 gpm/ft² (2.0 with polymer) |
| Thickening solids loading | 0.4–1.0 lb/hr/ft² (1.0–2.0 with polymer) |
| Thickening recycle | 100–200% |
| Thickening minimum feed | 5,000 mg/L |
| Thickened float concentration | 2–4% (3–5% with polymer) |
| Float blanket depth | 6–8 inches |
| Thickener subnatant target | Under 100 mg/L TSS |
A/S is an outcome, not a setting — your equipment divided by what the plant received that morning. Most DAF units that quietly stopped working had that number moved for them, by feed solids or by temperature.
This guide is built on the following sources. The thickening material and Table 11.2 are from the certification manual; design-equation figures are from vendor technical references and should be verified against a design text before sizing anything.
Worked bubble-physics and recycle examples in this guide are derived arithmetic from the stated inputs (Stokes’ law at 20°C; air release from Sa(fP−1); recycle by solving the A/S equation). They show why the two DAF processes differ by an order of magnitude — they are not design calculations, and turbidity/solids correlations are site-specific.
Rising sludge vs bulking sludge — the failure DAF copies · Solids thickening & dewatering · Nitrification & denitrification · Aeration systems — the other way air goes into water
Flotation, thickening, and solids-handling questions reward understanding over memorization. The Complete Exam Guide covers the whole span — 200 practice questions, written the way examiners write them.
Get the Complete Exam Guide — $17Instant PDF · One-time payment · Lifetime access