Question 3 of 5: Coagulation–Flocculation and Taste and Odour Control

Question 3: Coagulation–Flocculation and Taste and Odour Control (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.

(a) Coagulation and flocculation (15 marks)

The particles that make natural water turbid — clays, silts, algae, humic colloids — are in the size range 0.01 to 10 µm and carry a net negative surface charge at natural pH. Two facts follow. Their settling velocities are so small that plain sedimentation would take days to weeks, and their mutual electrostatic repulsion prevents them aggregating into anything that would settle faster. Coagulation and flocculation are the two-step answer: coagulation destabilises the colloid chemically in a violently mixed tank lasting seconds, and flocculation then grows the destabilised particles into settleable floc by gentle stirring over twenty to forty minutes. The distinction is important because the two steps want opposite mixing intensities, and applying the wrong one destroys the result.

The stability that must be overcome is described by DLVO theory. Each negative particle attracts a tightly held layer of counter-ions (the Stern layer) and, beyond it, a diffuse cloud that thins with distance — the electrical double layer. Two approaching particles feel van der Waals attraction, which is strong but short-ranged, and double-layer repulsion, which is weaker but longer-ranged; their sum has an energy barrier at intermediate separation. Particles with enough thermal energy to cross that barrier fall into the primary minimum and stick permanently. Coagulation works by lowering or removing the barrier, and the three mechanisms named in the question are three distinct ways of doing so.

Ionic layer compression is the most direct. Adding an indifferent electrolyte raises the ionic strength of the water, which compresses the diffuse layer — the Debye length varies as the inverse square root of ionic strength — bringing the plane of repulsion in close enough that van der Waals attraction dominates and the energy barrier vanishes. The effect is startlingly sensitive to counter-ion charge: the Schulze–Hardy rule holds that the coagulating concentration varies as the inverse sixth power of the counter-ion valence, so a trivalent ion such as Al³⁺ is several hundred times more effective than a monovalent one. This is the mechanism by which a river plume flocculates naturally where it enters the sea, but it is rarely used deliberately in water treatment because the salt doses required are impractically large.

Charge neutralisation is the mechanism that actually operates at normal coagulant doses. Hydrolysing metal salts — aluminium sulfate, ferric chloride, polyaluminium chloride — produce a family of positively charged hydrolysis products that adsorb specifically onto the negative particle surface, reducing the zeta potential toward zero. When the zeta potential falls into roughly ±5 mV the barrier disappears and collisions become productive. Because the mechanism is adsorption, it is stoichiometric: the optimum dose is proportional to the particle surface area present, and overdosing reverses the charge to positive and restabilises the colloid. That restabilisation zone is the signature of charge neutralisation and the reason a jar test shows a distinct turbidity minimum rather than a plateau. Cationic polymers work the same way, with the added mechanism of interparticle bridging when the chain is long enough to adsorb on two particles at once.

Sweep coagulation is what happens at higher doses, and is the mechanism most Canadian surface-water plants actually run on. When the coagulant dose exceeds the solubility of the metal hydroxide, a voluminous amorphous precipitate of Al(OH)₃ or Fe(OH)₃ forms throughout the water and settles, enmeshing colloidal particles mechanically as it goes. Sweep floc is robust: it does not restabilise on overdose, it works on low-turbidity water where there is too little surface area for charge neutralisation to find, and it removes natural organic matter by adsorption onto the fresh hydroxide surface. Its cost is a much larger volume of chemical sludge to handle.

Given. A plant treating 20 ML/d with a 15 kW rapid mixer in a 20 m³ chamber (30 s detention) followed by a 600 m³ flocculation basin driven at 0.6 kW for 30 min. Alum dose 30 mg/L; water at 20 °C so \(\mu = 1.002\times10^{-3}\ \text{Pa}\cdot\text{s}\).

Find. The velocity gradients in each stage, the dimensionless \(Gt\) products, and the alkalinity the alum will destroy.

Approach. Apply the Camp–Stein velocity gradient \(G = \sqrt{P/\mu V}\) to each stage and compare with the standard design windows, then take the alkalinity consumption from the alum hydrolysis stoichiometry.

1. Rapid-mix intensity. \[G = \sqrt{\frac{P}{\mu V}} = \sqrt{\frac{15\,000}{(1.002\times10^{-3})(20)}} = 865\ \text{s}^{-1},\] comfortably inside the 700–1000 s⁻¹ window that ensures the coagulant is dispersed before hydrolysis is complete. \(Gt = 865(30) \approx 2.6\times10^4\).
2. Flocculation intensity. The same relation with the much larger basin gives \[G = \sqrt{\frac{600}{(1.002\times10^{-3})(600)}} = 31.6\ \text{s}^{-1}, \qquad Gt = 31.6(1800) = 5.7\times10^4,\] so \(\boxed{\ Gt \approx 5.7\times10^{4},\ \text{within the } 10^4\text{–}10^5 \text{ design range}\ }\) Note the asymmetry: the flocculator carries thirty times the volume on one twenty-fifth of the power. High \(G\) here would shear the growing floc apart faster than it forms.
3. Alkalinity destroyed. From \(\text{Al}_2(\text{SO}_4)_3\cdot14\text{H}_2\text{O} + 6\,\text{HCO}_3^- \rightarrow 2\,\text{Al(OH)}_3 + 3\,\text{SO}_4^{2-} + 6\,\text{CO}_2 + 14\,\text{H}_2\text{O}\), each mole of alum (594.4 g) consumes six equivalents of alkalinity (300 g as CaCO₃), so a 30 mg/L dose destroys \(30(0.505) = 15.1\) mg/L as CaCO₃, releases 13.3 mg/L of CO₂, and produces 7.9 mg/L of Al(OH)₃ sludge.

A raw water with less than about 30 mg/L of alkalinity would therefore need lime or soda ash added with the alum, or the pH would collapse out of the 6.0–7.5 window in which aluminium hydroxide is least soluble — leaving residual dissolved aluminium in the finished water.

(b) Taste and odour (10 marks)

Definition. Taste and odour are the organoleptic properties of water — what it tastes and smells like to a human panel. They are not chemical analyses but sensory measurements, quantified either by the threshold odour number, the dilution factor at which the odour just becomes undetectable to a panel, or by flavour profile analysis, in which trained panellists score named descriptors (earthy, musty, chlorinous, fishy, septic) on an intensity scale. A sample of 25 mL diluted to 200 mL with odour-free water and just at threshold has a threshold odour number of \(200/25 = 8\). Taste and odour carry no health risk in themselves, but they are the property by which the public judges its water supply, and a musty episode generates more complaints than any other single event a utility faces — and drives consumers to less safe alternatives.

Typical sources. The dominant sources in Canadian surface supplies are the algal metabolites geosmin and 2-methylisoborneol, released by cyanobacteria and actinomycetes during and after a bloom; both are detectable by the human nose at 5–10 ng/L, roughly a thousand times below the concentration of anything else in the water. Others are decaying vegetation and humic material in peat-influenced water (earthy, woody); hydrogen sulfide from anoxic groundwater or a stratified reservoir hypolimnion (rotten egg); iron and manganese (metallic, bitter); industrial solvents and phenols, which chlorine converts to intensely odorous chlorophenols; chlorine and chloramine residuals themselves; and, in distribution, biofilm growth and stagnant dead ends.

Removal processes. Two are described here.

Adsorption on activated carbon. Powdered activated carbon dosed ahead of the rapid mix, or a granular activated carbon contactor, removes geosmin and 2-methylisoborneol by physical adsorption into a pore structure with an internal surface area of order 1000 m²/g. Equilibrium follows the Freundlich isotherm \(q_e = K_fC_e^{1/n}\), so the required dose is obtained from a mass balance between the influent and target concentrations. Powdered carbon suits seasonal, bloom-driven episodes because it can be dosed only when needed; granular carbon suits a chronic problem but competes with natural organic matter for sites and needs periodic reactivation.

Chemical oxidation. Ozone, or an advanced oxidation process combining ozone with hydrogen peroxide or ultraviolet light, destroys the odorant rather than transferring it to another phase. Geosmin and 2-methylisoborneol are saturated alicyclic alcohols and resist molecular ozone, so the effective agent is the hydroxyl radical, which is why advanced oxidation outperforms ozone alone on exactly these two compounds. Chlorine and permanganate are effective on sulfide and on iron- and manganese-related odours but not on the algal metabolites; chlorine applied to a phenolic water makes the problem markedly worse.

Two further options deserve a mention: aeration, which strips volatile odorants such as hydrogen sulfide and some solvents but does nothing for the non-volatile algal metabolites; and biological filtration, in which an unchlorinated granular filter develops a biofilm capable of metabolising geosmin, an increasingly common Canadian retrofit.

Given. A reservoir bloom raising geosmin to 40 ng/L. Powdered activated carbon with a Freundlich isotherm \(K_f = 1.2\) and \(1/n = 0.45\) in units of ng/mg and L/ng; the target is 8 ng/L, below the odour threshold.

Find. The powdered-carbon dose required.

1. Equilibrium loading. At the target residual, \(q_e = K_fC_e^{1/n} = 1.2(8)^{0.45} = 3.06\ \text{ng/mg}\).
2. Mass balance. The carbon must carry the mass removed, so \[m = \frac{C_0 - C_e}{q_e} = \frac{40 - 8}{3.06} \;\Longrightarrow\; \boxed{\ m = 10.5\ \text{mg/L of PAC}\ }\] a dose squarely inside the practical 5–20 mg/L range, and one that would be applied only for the few weeks of the bloom.