Coagulation and Flocculation: Colloid Science to Design

What is the difference between coagulation and flocculation?

Coagulation is the chemical destabilisation of colloids — you add a coagulant that neutralises or screens the surface charge that keeps particles apart, so they stop repelling one another. Flocculation is the subsequent physical transport step in which slow, controlled mixing brings the destabilised particles into contact so they grow into aggregates (flocs) large enough to remove by sedimentation, flotation or filtration.

The distinction matters because the two steps have opposite mixing requirements. Coagulation needs a brief, intense rapid mix to disperse the coagulant before it hydrolyses; flocculation needs prolonged gentle agitation that promotes particle collisions without shearing the growing flocs apart. Confusing the two — over-mixing the floc, or under-dispersing the coagulant — is the single most common cause of poor clarifier or dissolved-air flotation performance.

Why are colloids stable? Double layer, zeta potential and DLVO

Colloidal particles (roughly 1 nm–1 µm: clays, bacteria, emulsified oil, natural organic matter) carry a net surface charge, almost always negative at natural pH. That charge attracts a cloud of counter-ions, forming the electrical double layer: a tightly bound Stern layer plus a diffuse layer that decays over the Debye length, 1/κ. The potential measured at the hydrodynamic slip plane is the zeta potential (ζ) — the practical, measurable proxy for particle charge and therefore for stability.

Whether two approaching particles aggregate is governed by DLVO theory (Derjaguin–Landau–Verwey–Overbeek), which sums two opposing energies as a function of separation distance H:

Van der Waals attraction scales as 1/H and is essentially fixed by the material; electrostatic repulsion scales with ζ² and decays exponentially over the Debye length. Coagulation works by attacking the repulsive term — reducing ζ toward zero, or compressing 1/κ — until the energy barrier collapses and every collision sticks. In practice, destabilisation is achieved once |ζ| falls into roughly the −10 to +5 mV window; the aim is not charge reversal but charge cancellation.

What are the four coagulation mechanisms?

Four distinct mechanisms can destabilise a colloid, and a real coagulant dose usually recruits more than one:

1. Double-layer compression. Adding indifferent electrolyte raises the ionic strength, which shortens the Debye length (1/κ ∝ I^−1/2) and squeezes the diffuse layer. The repulsive barrier shrinks even though ζ is unchanged. This is the mechanism behind the Schulze–Hardy rule, where coagulating power rises steeply with counter-ion valence (Na⁺ « Ca²⁺ « Al³⁺).
2. Adsorption and charge neutralisation. Positively charged metal-hydrolysis species or cationic polymers adsorb directly onto the negative colloid and cancel its charge. This dominates at low coagulant doses; overdosing reverses the charge and re-stabilises the suspension (restabilisation).
3. Sweep flocculation (enmeshment). At higher metal-salt doses and near-neutral pH, the coagulant precipitates as a bulky amorphous metal hydroxide, Al(OH)₃ or Fe(OH)₃, which physically enmeshes colloids as it settles. This is the workhorse mechanism for most water and wastewater plants.
4. Interparticle bridging. Long-chain polymers adsorb at points onto several particles at once, tethering them into a network. Bridging produces large, strong, shear-resistant flocs — but excess polymer or excessive shear causes steric restabilisation.

How do metal-salt coagulants work, and why do they consume alkalinity?

Aluminium (alum) and iron (ferric) salts do not act as simple ions. On contact with water the metal cation hydrolyses through a cascade of species — for aluminium, Al³⁺ → Al(OH)²⁺ → Al(OH)₂⁺ → polynuclear species such as Al₁₃ → amorphous Al(OH)_3(s). The distribution of these species is strongly pH-dependent, which is why every coagulant has an optimum pH window (roughly pH 6.0–7.5 for alum, 5–8.5 for ferric). Each hydrolysis step releases protons, so the reaction is acidic and consumes alkalinity:

Worked alkalinity check: dosing 60 mg/L alum consumes 60 × 0.50 = 30 mg/L alkalinity as CaCO₃. If the raw water carries only 40 mg/L, the residual 10 mg/L is too little to buffer the pH, which will crash out of the optimum window and stall hydroxide precipitation. The fix is to add supplemental alkalinity — lime or sodium bicarbonate — to restore buffering. Ignoring this is a classic reason a jar test looks fine but the full-scale plant underperforms once the natural buffer is exhausted.

What do polymer flocculants add?

Synthetic polyelectrolytes (typically polyacrylamide-based, supplied anionic, cationic or non-ionic) are used either as primary coagulants — where cationic polymers neutralise charge directly — or, far more often, as flocculant aids dosed after the metal salt. Their value is the bridging mechanism: a high-molecular-weight chain adsorbs onto multiple destabilised particles and knits them into large, dense, shear-resistant flocs that settle faster and dewater better.

Because bridging depends on chains spanning the gap between particles, polymer dosing is unforgiving of both under- and over-dose. Too little polymer leaves particles untethered; too much coats every surface so no bare sites remain for a chain to bridge to, causing steric restabilisation. Polymer is also shear-sensitive — once a bridged floc is broken by excessive G it does not fully re-form, so aids are dosed at the end of coagulation or early flocculation and never subjected to a second rapid mix.

Rapid mix vs flocculation: the physics of G and Gt

The mixing intensity for each stage is quantified by the velocity gradient G (the Camp–Stein root-mean-square shear rate), which sets the rate of particle collisions in orthokinetic (shear-driven) flocculation:

Worked example — rapid mix G: a flash mixer dissipates P = 500 W into a V = 1 m³ chamber of water at 15°C (μ = 1.14×10⁻³ Pa·s). Then:

• G = √(500 / (1.14×10⁻³ × 1)) = √(438,600) ≈ 662 s⁻¹.
• At a rapid-mix residence time t = 30 s, the Camp number is Gt = 662 × 30 ≈ 19,900.

A G of ~660 s⁻¹ sits squarely in the rapid-mix band (250–1,500 s⁻¹), enough to disperse the coagulant in seconds before hydrolysis completes. The same intensity applied to flocculation would tear the flocs apart — flocculation targets G of only 10–80 s⁻¹ over 15–45 minutes. Note that G falls with rising temperature (viscosity drops), so a mixer delivering 662 s⁻¹ in winter delivers more in summer for the same power draw, which is why tapered flocculation and variable-speed drives are specified. To size the physical mixing and separation train around these numbers, see our guide on selecting water treatment equipment.

Typical G, Gt and coagulant values

The table below gives defensible starting ranges for design and jar-test interpretation. Treat them as first-pass values to be confirmed on the actual water.

Pre-hydrolysed coagulants such as PACl carry their own base, so they consume far less alkalinity and work over a wider pH range — a practical advantage on soft, poorly buffered waters.

How is the optimum dose and pH found in practice?

No amount of theory replaces the jar test. Colloid surface chemistry, natural organic matter and alkalinity vary too much between waters to predict a dose from first principles — DLVO and the four mechanisms tell you why a dose works, not what the number is. The jar test brackets the coagulant type, dose and pH; a streaming-current or zeta-potential meter then tightens the charge-neutralisation optimum, and settled-water turbidity or a downstream flotation test confirms floc quality.

Once the chemistry is fixed, the separation stage decides the rest. Well-conditioned, dense flocs suit sedimentation; light or oily flocs are better lifted by flotation. For fats, oils and emulsified solids, coagulation and flocculation are almost always paired with dissolved-air flotation — see what a DAF system is and how to size a DAF system to match the chemistry to the hardware.

How to develop a coagulation and flocculation programme

1. Characterise the water. Measure turbidity, TSS, pH, alkalinity, temperature and target contaminants. Alkalinity and pH set which coagulant is viable and whether supplemental base is needed.
2. Screen coagulants by jar test. Run parallel jars across a dose range for each candidate (alum, ferric, PACl). Use a rapid mix of ~1 min at high G, then slow mix, then settle.
3. Optimise dose and pH together. Vary pH at the best dose to find the joint optimum; confirm with zeta potential or streaming current near the destabilisation window (|zeta| < ~10 mV).
4. Add a flocculant aid if needed. Trial a low-dose polymer after the metal salt to build larger, shear-resistant flocs; watch for restabilisation from overdose.
5. Set mixing energy. Specify rapid-mix G of 250-1,500 s-1 for seconds and flocculation G of 10-80 s-1 for 15-45 min, using G = sqrt(P/muV) to size the drives.
6. Confirm on the separation stage. Validate the chemistry against the actual clarifier, filter or DAF, and re-check alkalinity consumption at full-scale dose.