Question 4 of 5: Coagulation Mechanisms; Requirements of a Water Distribution System

Question 4: Coagulation Mechanisms; Requirements of a Water Distribution System (25 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Part (a) — Inter-particle bridging, sweep coagulation and ionic layer compression (12 marks).

Given. Colloidal clay and natural organic matter in a surface water, negatively charged at the pH of treatment; alum dose 30 mg/L as Al₂(SO₄)₃·14H₂O; ionic strength adjustable between 10⁻³ and 10⁻¹ M. Find. The physical basis of each of the three destabilisation mechanisms, and the conditions under which each dominates.

Colloids do not settle because they are small and because they are stable. Smallness is a kinetic problem — a 1 μm clay particle settles at less than 1 m/d — but stability is the electrical one: each particle carries a surface charge, surrounded by a compact Stern layer and a diffuse layer of counter-ions, and when two particles approach, the overlap of their diffuse layers produces a repulsion that, in DLVO terms, raises an energy barrier above the van der Waals attraction. Every coagulation mechanism is a way of removing that barrier or bypassing it.

1. Ionic layer compression (double-layer compression).

   Adding an indifferent electrolyte — one whose ions do not adsorb specifically — raises the ionic strength and packs the counter-ions closer to the surface. The characteristic thickness of the diffuse layer is the Debye length, which for a symmetrical electrolyte at 25 °C is

   $$\kappa^{-1} = \frac{0.304}{\sqrt{I}}\ \mathrm{nm}\quad (I\ \mathrm{in\ mol/L})$$

   so raising the ionic strength a hundredfold from 10⁻³ to 10⁻¹ M shrinks it by a factor of ten:

   $$\boxed{\ \kappa^{-1}: 9.6\ \mathrm{nm} \rightarrow 0.96\ \mathrm{nm}\ }$$

   Once the repulsive layer is thinner than the range of the van der Waals attraction, the energy barrier vanishes and every collision sticks. The controlling variable is counter-ion charge, not dose: the Schulze–Hardy rule makes the critical coagulation concentration scale as z⁻⁶, so relative to Na⁺ the required molar concentration falls by 2⁶ = 64 for Ca²⁺ and by 3⁶ = 729 for Al³⁺. This is the mechanism that explains why a river plume flocculates on meeting the sea, but it is not the mechanism used in a water treatment plant — the salt concentrations required are far too high to leave a potable water.

2. Adsorption and inter-particle bridging.

   A long-chain polymer — a synthetic polyacrylamide, or a hydrolysed aluminium polymer — adsorbs onto a particle at a few segments while the remainder of the chain extends into solution as loops and tails. A tail that reaches a vacant site on a second particle binds the two together, forming a physical bridge whose strength does not depend on charge neutralisation at all. This is why non-ionic and even anionic polymers can flocculate negatively charged clay.

   Two features follow, and both are practical. First, bridging produces the strongest and largest flocs, which is why polymer is added as a flocculant aid ahead of filtration and as a conditioner ahead of sludge dewatering. Second, the dose has a sharp optimum: too little polymer leaves too few bridges, while too much covers every available site so that no vacant surface remains for a tail to attach to. That overdose condition is restabilisation, and it also arises through a purely electrical route with cationic polymers — charge reversal past the isoelectric point. Bridging also requires the right mixing: the chain must be given time to extend before the next collision, so a slow, prolonged flocculation follows a very brief dispersion.

3. Sweep coagulation (enmeshment in a precipitate).

   When a hydrolysing metal salt is dosed well above its solubility limit, the dominant product is not a charge-neutralising species but a bulk amorphous hydroxide precipitate, Al(OH)_3(am) or Fe(OH)_3(am), which forms throughout the water and settles, physically enmeshing the colloids it encounters on the way down. With alum consuming natural alkalinity,

   $$\mathrm{Al}_2(\mathrm{SO}_4)_3\!\cdot\!14\mathrm{H}_2\mathrm{O} + 6\,\mathrm{HCO}_3^{-} \rightarrow 2\,\mathrm{Al}(\mathrm{OH})_3(s) + 3\,\mathrm{SO}_4^{2-} + 6\,\mathrm{CO}_2 + 14\,\mathrm{H}_2\mathrm{O}$$

   Every milligram of alum consumes 0.505 mg/L of alkalinity as CaCO₃ and yields 0.262 mg/L of hydroxide floc, so at 30 mg/L,

   $$\boxed{\ \text{alkalinity consumed} = 15.2\ \mathrm{mg/L\ as\ CaCO_3};\quad \mathrm{Al(OH)_3} \ \text{formed} = 7.9\ \mathrm{mg/L}\ }$$

   Sweep coagulation is the workhorse mechanism of conventional Canadian water treatment, and its appeal is that it is robust: it does not require a precisely matched dose, it works on low-turbidity waters where collisions between colloids would otherwise be too rare, and the removal rate rises with dose rather than passing through a restabilisation window. Its costs are the alkalinity consumption — which must be replaced with lime or soda ash in a soft water, or the pH will fall out of the 6.0–7.8 window where Al(OH)₃ is least soluble — and a large volume of chemical sludge.

In a real jar test all three appear on one curve. At low dose, charge neutralisation and layer compression give a narrow optimum; raise the dose and the water restabilises; raise it further, past the hydroxide solubility limit, and the sweep zone opens up and stays open. A plant chooses the sweep zone for reliability and adds a bridging polymer to give the resulting floc the strength it otherwise lacks.

Part (b) — Requirements of an adequate distribution system; grid-iron versus dead-end (13 marks).

The key requirements of an adequate water distribution system are:

• Quantity. Deliver the maximum-day demand plus the fire flow simultaneously, and the maximum-hour demand on its own — whichever governs. In Canadian practice fire flow is sized by the Fire Underwriters Survey method and is typically the governing case for main sizing in residential areas.
• Pressure. Maintain roughly 275–415 kPa (40–60 psi) under normal demand, with a minimum of about 140 kPa (20 psi) at every point during a fire flow, and a maximum below about 700 kPa to limit leakage and fixture damage. Positive pressure at all times is also a water-quality requirement, since it is the barrier against intrusion.
• Water quality preserved to the tap. A disinfectant residual maintained throughout, water age kept short enough to avoid nitrification and disinfection-by-product formation, no dead ends where stagnation occurs, corrosion control, and physical protection against cross-connections and backflow.
• Reliability and redundancy. Two-directional supply to critical areas, valving arranged so that a break isolates a short reach rather than a district, and adequate storage — equalisation, fire reserve and emergency reserve — positioned to float on the system.
• Durability and maintainability. Materials and bedding suited to the ground and the frost depth (a real constraint in Canada: mains below the local frost line, commonly 1.8–2.4 m), hydrants, valves, blow-offs and air-release valves accessible, and the whole system mappable and repairable without excessive shutdown.
• Economy. Least total cost over the design life — capital plus pumping energy plus maintenance — while meeting the criteria above.

Grid-iron (looped) system. Mains are laid on a rectangular pattern and interconnected, so every node is fed from two or more directions. Its advantages: no dead ends, so no stagnation and a uniform disinfectant residual; water reaches any point by several paths, so a break can be isolated by closing a few valves without cutting off more than a block; and the divided flow means much lower head loss and better pressure during fire flow. That last advantage is quantitative and is worth demonstrating. For a 150 L/s fire flow reaching a hydrant 1200 m from the source through 300 mm main (C = 120), the Hazen–Williams loss is

In a dead-end layout one main carries the whole 150 L/s; in a loop the same demand arrives along two comparable paths, each carrying half. Since head loss varies as Q^1.852, halving the flow in each leg gives

Its disadvantages: more pipe length and more valves for the same served area, so higher capital cost; the flow distribution is indeterminate and requires a network analysis (Hardy Cross, or a modern solver) rather than a simple series calculation; and more interconnections mean more fittings to leak and a more complex flushing programme.

Dead-end (tree or branching) system. A main trunk branches into sub-mains and then into branches that terminate. Its advantages: it is the cheapest layout for a given served area, uses the least pipe, is simple to lay out and to analyse because the flow in every pipe is uniquely determined by the demands beyond it, and pipe sizes taper naturally down the branch. It suits ribbon development along a highway, or a rural extension, where a loop would have nothing to loop back to. Its disadvantages are severe in an urban setting: a break or a valve closure anywhere on a branch cuts off everything downstream; water stagnates at the extremities, so the residual decays, sediment accumulates and taste, odour and coliform problems develop, requiring a programmed flushing regime that wastes water; the available fire flow at a dead end is limited by the single feed; and pressure at the far end is the lowest in the system, so a peak demand anywhere upstream is felt there first.

In practice the choice is not binary. Canadian municipal standards require looping of all mains serving hydrants wherever it is practicable, permit a temporary dead end at the leading edge of a development on the condition that it be looped when the adjacent phase is built, and require an automatic flushing device or a regular flushing programme wherever a permanent dead end is unavoidable.