Coagulants, polymers and jar testing for DAF

A dissolved air flotation unit is a separator. The chemistry decides what there is to separate. Get the coagulant and the pH right and a mediocre machine performs well; get them wrong and the best machine on the market will not save you.

Direct answer

Dissolved air flotation needs a coagulant to neutralise particle charge and a flocculant polymer to bind the neutralised particles into flocs that bubbles can lift. Polyaluminium chloride at 50–600 mg/L is the most widely applicable coagulant because it works across pH 6.0–8.5 and produces less sludge than alum. Ferric chloride at 100–800 mg/L is preferred where phosphorus or sulphide must also be removed, and on tannery and high-ionic-strength streams. Anionic polyacrylamide at 1–5 mg/L suits most food, oil and municipal duties; cationic polymer at 0.5–3 mg/L is correct for paper mill whitewater. Doses must be confirmed by jar test on a real composite sample — published ranges are starting points, not design values.

Coagulant selection

CoagulantWorking pHDose rangeStrengthsWeaknesses
Polyaluminium chloride (PAC)6.0 – 8.550 – 600 mg/LWide pH window, low alkalinity consumption, 30–50 % less sludge than alum, fast flocCosts more per kg; aluminium residual matters for potable duty
Aluminium sulphate (alum)5.8 – 7.5100 – 500 mg/LCheap and widely available across AfricaNarrow pH window, consumes alkalinity, more sludge, slow at low temperature
Ferric chloride4.5 – 6.5 and >8.5100 – 800 mg/LDense strong floc, removes phosphorus and sulphide, works at high ionic strengthCorrosive, stains, colours the float, high alkalinity demand
Ferric sulphate4.0 – 6.0 and >8.5100 – 700 mg/LLess corrosive than ferric chloride, no chloride addedAdds sulphate, which matters where anaerobic treatment follows
Cationic organic coagulant (polyDADMAC)4 – 1010 – 150 mg/LAlmost no added sludge, no metal residual, wide pH toleranceCosts considerably more; less effective alone on heavy solids
Lime>9.5VariableCheap; precipitates phosphorus, metals and some hardnessVery high sludge volume; handling and scaling problems
In much of Africa, availability and lead time constrain this choice more than technical merit. Confirm the local supply chain before designing around a coagulant your site cannot reliably buy — a plant designed for PAC and run on whatever arrives is a plant that will not perform.

Polymer selection

ChargeUse onDoseMolecular weight
AnionicFood, dairy, abattoir, edible oil, municipal, after metal-salt coagulation1 – 5 mg/LHigh to very high
CationicPaper mill whitewater, sludge thickening, streams with heavy anionic trash0.5 – 3 mg/LMedium to high
Non-ionicRare; very low or very high pH where ionic polymers hydrolyse1 – 4 mg/LHigh
Rule of thumb: after a metal-salt coagulant has neutralised charge, the floc surface is close to neutral or slightly positive, so anionic polymer bridges it best. Where no metal coagulant is used and the particles remain strongly negative — paper furnish is the classic case — cationic polymer is correct.

Starting doses by industry

IndustryCoagulant and doseTarget pHPolymer
Abattoir / meatFeCl₃ 100 – 400 mg/L or PAC 150 – 4006.0 – 7.0Anionic 1 – 4 mg/L
Edible oilPAC 150 – 600 mg/L, after acid split6.5 – 8.0Anionic 2 – 5 mg/L
DairyPAC 80 – 300 mg/L6.5 – 7.5Anionic 1 – 3 mg/L
BreweryPAC 50 – 200 mg/L6.5 – 7.5Anionic 1 – 3 mg/L
PoultryPAC 120 – 350 mg/L6.0 – 7.0Anionic 1 – 4 mg/L
TanneryFeCl₃ 200 – 800 mg/L7.5 – 8.5Anionic 2 – 6 mg/L
Pulp & paperPAC 50 – 250 mg/L or none6.5 – 8.0Cationic 0.5 – 3 mg/L
Textile dyeingPAC 150 – 500 mg/L + decolourant6.0 – 8.0Cationic 1 – 4 mg/L
MunicipalFeCl₃ 30 – 120 mg/L or alum 40 – 1506.0 – 7.5Anionic 0.5 – 2 mg/L
Surface water / algaePAC 10 – 60 mg/L6.0 – 7.5Often none

A jar test that predicts full-scale performance

Most jar tests fail to predict plant behaviour because they replicate settling, not flotation, and because the mixing regime bears no relation to the machine. This protocol addresses both.

  1. Sample properly. Composite over a full production cycle, including CIP and shift changeover. Test within 6 hours or refrigerate; a sample left overnight at African ambient has already changed.
  2. Measure the baseline. pH, temperature, TSS, total COD and filtered COD (0.45 µm). The filtered COD sets the ceiling on what flotation can achieve, before any chemistry is chosen.
  3. Set pH first. Adjust each jar to the target for the coagulant under test. Testing coagulant dose at the wrong pH tells you nothing.
  4. Rapid mix. Add coagulant at 200 rpm for 30 seconds. This approximates G ≈ 500 s⁻¹.
  5. Slow mix with polymer. Reduce to 40 rpm, dose polymer, continue 10 minutes. Approximates G ≈ 40 s⁻¹. Watch when the floc first becomes visible — that timing informs flocculation chamber sizing.
  6. Simulate flotation, not settling. This is the step usually omitted. Use a bench flotation cell, or inject 10–20 % by volume of water previously saturated at 5 bar in a small pressure vessel. Stop stirring and observe for 5 minutes.
  7. Assess. Record float thickness and firmness, clarity of the subnatant, whether any floc sinks, and how the float behaves when disturbed — a float that collapses when nudged will collapse under a skimmer.
  8. Analyse the subnatant. TSS, FOG, total COD. These figures, not the visual appearance, are the result.
  9. Bracket the optimum. Run at least five doses spanning half to twice the expected value. The optimum is a range, not a point, and the width of that range tells you how much control the plant will need.

Reading the result

ObservationLikely causeAction
No visible flocUnderdosed coagulant, or pH outside windowCheck pH first, then increase coagulant
Fine floc that will not growInsufficient polymer, or polymer shearedIncrease polymer; reduce slow-mix speed
Good floc that sinksOverdosed metal coagulant; floc too denseReduce coagulant, increase polymer
Slimy float that will not firm upOverdosed polymer, re-stabilisationReduce polymer by half and re-test
Clear subnatant, thin floatInsufficient air, not a chemistry problemIncrease simulated recycle; check plant A/S ratio
Cloudy subnatant, good floatCarry-over of unflocculated finesExtend flocculation time; check for short-circuiting

Frequently asked questions

Which coagulant is best for DAF?

Polyaluminium chloride for most industrial streams, because it works across pH 6.0–8.5, forms floc quickly, consumes little alkalinity and produces 30–50 % less sludge than alum. Ferric chloride is better where phosphorus or sulphide must also be removed, where ionic strength is high (tannery, produced water), or where a denser, more shear-resistant floc is needed. Alum remains a reasonable choice where price and local availability dominate and pH is stable.

How do I calculate coagulant dose?

You do not calculate it — you measure it. Coagulant demand depends on particle charge, natural organic matter, alkalinity and ionic strength in combinations that no formula predicts reliably. Use the published range for your industry as a starting point, bracket it with a jar test across at least five doses, then confirm at plant scale during commissioning. Recheck whenever the process, the season or the coagulant supplier changes.

Can a DAF run without chemicals?

Only where the target material is already buoyant and unstabilised: free oil, coarse fibre, some algae. Anything emulsified, colloidal or charge-stabilised will pass straight through an unchemically-dosed flotation cell, because the particles are too small for a bubble to attach to usefully. If reliable chemical supply is genuinely doubtful at your site, that constraint should shape the process selection at the start, not surface as a complaint later.

Why does my jar test look good but the plant performs badly?

Four usual reasons. The jar test simulated settling rather than flotation, so it never tested whether the floc floats. The mixing energies did not match the plant, so floc that survived a paddle is shearing in a pump or nozzle. The sample was a grab rather than a composite, so it did not represent what the plant actually receives. Or the plant’s dosing is not delivering what the setpoint claims — measure the pump output with a calibration column before blaming the chemistry.