In atmospheric pressure chlorine absorption towers fed with sodium hydroxide solution, the stoichiometric basis for caustic feed is normally derived from the net absorption reaction Cl₂ + 2 NaOH → NaOCl + NaCl + H₂O. The theoretical demand is 1.128 kg NaOH per 1 kg Cl₂, corresponding to 2.256 kg of 50 wt% caustic solution per 1 kg Cl₂. In continuous sodium hypochlorite production this value is not a stable operating setpoint because the scrubbing liquor must retain a measurable excess of free sodium hydroxide to keep hypochlorous acid concentration low and suppress chlorine desorption from the sump liquor. A single-stage packed tower producing commercial sodium hypochlorite may operate at a feed molar ratio of 2.05 to 2.15 mol NaOH per mol Cl₂, with excess caustic in the recirculation loop maintained at 0.5 wt% to 2.0 wt% NaOH. The tower itself is usually a countercurrent packed column with random packing such as 25 mm or 50 mm polypropylene Pall rings, operating at superficial gas velocities between 0.6 m s⁻¹ and 1.5 m s⁻¹. Under these conditions, the liquid-side reaction between dissolved chlorine and hydroxyl ion is sufficiently rapid that the dominant mass-transfer resistance is in the gas film, and the apparent scrubbing efficiency is governed by the availability of free caustic at the gas-liquid interface. The caustic feed system is therefore designed as a feed-forward loop in which the flow of sodium hydroxide solution is ratioed to the chlorine gas mass flow, with a pH-based trim signal from the recirculating liquor. Failure to maintain this excess caustic produces a sharp loss of absorption efficiency because the equilibrium partial pressure of chlorine above the liquor rises rapidly when the pH falls below about 10.5.
When a bleach plant scale-up transfers a laboratory titration curve directly to a single-stage packed tower without accounting for the vertical gradient in gas-phase chlorine concentration, the observed caustic demand is typically 3% to 7% higher than the theoretical 1.128 kg kg⁻¹ because the equilibrium partial pressure of chlorine above the spent scrubbing liquor is governed by hypochlorous acid hydrolysis, not by the bulk free sodium hydroxide concentration alone. At 25°C, hypochlorous acid has a dissociation constant of approximately 7.54 on the pH scale. At pH 10.5, the ratio of hypochlorite ion to hypochlorous acid is 912, so the undissociated acid fraction is 0.11% of total free available chlorine. At pH 9.0, that ratio falls to 28.8 and the undissociated acid fraction rises to 3.4%. Because hypochlorous acid possesses a materially higher vapor pressure than the hypochlorite ion, a fall in sump pH from 10.5 to 9.0 can increase chlorine vapor losses by more than an order of magnitude even when residual NaOH is still measurable by conventional titration. In a tower that processes chlorine gas containing traces of hydrochloric acid or chlorinated organic compounds, the pH excursion is usually localized near the top of the packing where the chlorine partial pressure is lowest and the liquid flow is furthest from the caustic injection point. This is why the feed-forward molar ratio must be biased upward when the top-of-bed pH transmitter and the sump pH transmitter diverge by more than 0.3 pH units.
The apparent breakdown of the 2:1 molar relationship occurs not because the net absorption reaction changes, but because the tower must neutralize competing acidic species and compensate for slow degradation of sodium hypochlorite in the recirculating liquor. Sodium hypochlorite decomposes along two important routes: oxygen evolution and chlorate formation. The chlorate route, 3 NaOCl → NaClO₃ + 2 NaCl, is second-order with respect to hypochlorite ion; therefore an increase in product strength from 10 wt% to 15 wt% NaOCl raises the kinetic rate by a factor of 2.25 if temperature and ionic strength remain unchanged. The decomposition of hypochlorite ion reduces the concentration of the alkaline hydrolysis species and shifts the equilibrium Cl₂ + H₂O ⇌ HOCl + H⁺ + Cl⁻ toward the acid side, thereby consuming part of the caustic inventory that the feed-forward controller assumes to be available for fresh chlorine absorption. In practice, a compensated caustic feed system calculates the required sodium hydroxide molar flow as the sum of 2.00 mol NaOH per mol Cl₂, plus 2.00 mol NaOH per mol CO₂, plus 1.00 mol NaOH per mol HCl, plus an adjustable excess term for decomposition and pH trim. For a chlorine feed of 12.7 kg h⁻¹, the theoretical NaOH demand is 14.3 kg h⁻¹; at a NaOH to Cl₂ ratio of 2.10, the demand becomes 15.1 kg h⁻¹, and with 20 wt% caustic the required solution flow is 75.5 kg h⁻¹. The corresponding caustic feed pump should be selected with a turndown of at least 10:1 to accommodate low-flow trim when chlorine demand swings during cell-room load changes.
Temperature excursions above 25°C are the most common cause of apparent stoichiometric drift in a continuous bleach tower. Published kinetic data for concentrated sodium hypochlorite indicate that chlorate formation accelerates by a factor of 2 to 3 for each 10°C rise in the 15–35°C range, although the exact factor depends on chloride concentration, ionic strength, and trace-metal contamination. A plate-and-frame heat exchanger with chilled water at 5°C to 10°C is typically installed in the recirculation line to hold the sump temperature between 18°C and 22°C when the chlorine absorption load exceeds 40 kW of exothermic release. The heat release from chlorine absorption is approximately 100 kJ per mol Cl₂, so a chlorine feed of 12.7 kg h⁻¹ releases roughly 17.9 MJ h⁻¹, or 5.0 kW. That heat load does not appear excessive, but when the recirculation loop is small and the tower is exposed to ambient solar gain, sump temperatures can still rise above 25°C and invalidate the excess caustic setpoint. The processing window is narrow: a sustained temperature deviation of only ±5°C can shift the required free-caustic concentration by several tenths of a percentage point because the hydrolysis constant of hypochlorous acid is temperature-dependent and because the rate of chlorate formation increases nonlinearly.
Caustic feed stoichiometry cannot be separated from liquid distribution in a packed tower because the free sodium hydroxide must reach the same gas-liquid interfacial area that the chlorine gas contacts. A tower treating 1,000 m³ h⁻¹ of scrubber off-gas at 101.325 kPa and 20°C with a superficial gas velocity of 1.0 m s⁻¹ requires a cross-sectional area of 0.278 m², equivalent to a column diameter of 0.595 m. For a liquid loading of 12 m³ m⁻² h⁻¹, the recirculation flow is 3.3 m³ h⁻¹, and the liquid distribution system must deliver that flow uniformly across the packing surface. If caustic is introduced as 50 wt% sodium hydroxide directly into the top distributor, the high-density caustic stream forms localized rivulets that do not mix rapidly with the recirculating bleach; the local pH can exceed 13 within those rivulets, causing carbonate and silicate precipitation and reducing the effective packing area. Dilution to 20 wt% caustic before injection is therefore standard practice, with the dilution water softened to a total hardness of <17 mg/L as CaCO₃. In addition, the caustic feed nozzle should be placed in the suction side of the recirculation pump or into a static mixer upstream of the distributor to avoid density-driven segregation. The pH probe that trims the caustic feed is normally installed in a flowing bypass loop on the recirculation line with a sample velocity of 0.3 m s⁻¹ to 0.6 m s⁻¹, because stagnant pH measurements in the sump introduce a dead time of 2 to 3 minutes that destabilizes the ratio loop.
Hydraulic limits also constrain how much excess caustic can be carried before the tower loses absorption performance. At superficial gas velocities above 1.5 m s⁻¹ with random packing, liquid entrainment increases and the mist eliminator becomes a source of caustic carryover into the stack. At superficial gas velocities below 0.6 m s⁻¹, wetting is inadequate and gas channels through dry packing; in such a case the actual molar ratio at the interface may be locally below 2.0 even though the global feed ratio is 2.10. The caustic feed control system should therefore include a constraint that suspends the excess-alkalinity increase when the recirculation flow is below the minimum wetting rate for the selected packing. Manufacturers’ technical bulletins for random packing typically define a minimum liquid distributor drip density of 50 to 70 points per m² for random and structured packings with nominal sizes of 25 mm to 50 mm; published data for this specific chlorine-reactive system is limited, but the general wetting limits are broadly applied. The operational boundary is that the pump turndown and the liquid distributor design together determine the lowest chlorine load at which stoichiometric control can be maintained without localized pH collapse.
Emergency chlorine scrubber feed systems are rarely operated at steady-state stoichiometric ratio because the design chlorine release rate for a cylinder storage room is based on a presumed loss-of-containment event rather than on normal vent flow. A 68 kg chlorine cylinder requires 76.7 kg sodium hydroxide for complete conversion to sodium hypochlorite and salt; a 907 kg ton container requires 1,023 kg NaOH. If the scrubber sump holds 20 wt% caustic at a specific gravity of 1.22, the complete-release inventory for a ton container is approximately 5,115 kg of solution, occupying 4.2 m³. In practice, the scrubber sump is charged with a fixed sodium hydroxide inventory and the circulation pump is sized to provide the liquid flow required to meet the expected gas flow and chlorine concentration, while the caustic feed pump is actuated not by a chlorine flow signal but by a pH controller that adds fresh 20 wt% to 25 wt% caustic when the sump pH falls below a setpoint of 11.5. The emergency tower may operate only a few hours per year, so the pH probe and feed valve must be automatically exercised; a fast-acting control valve with a 2 s to 3 s stroke time is preferred because caustic demand can rise from background to maximum within the time required for a cylinder valve to fail open. Chlorine Institute Pamphlet 86 provides design guidance for such systems, and the discharge stack should be designed to meet an outlet chlorine concentration below the OSHA 29 CFR 1910.1000 Table Z-1 ceiling of 1 ppm and the NIOSH short-term exposure limit of 0.5 ppm over 15 min.
Carbon dioxide enters an atmospheric pressure chlorine absorption tower either as an impurity in the chlorine gas or as a component of air drawn into the tower when the gas train operates under slight negative pressure. Each mole of CO₂ consumes 2 mol of NaOH by the reaction CO₂ + 2 NaOH → Na₂CO₃ + H₂O, and the mass demand is 1.818 kg NaOH per 1 kg CO₂. This is a larger mass ratio than the chlorine demand of 1.128 kg kg⁻¹ because the molar mass of CO₂ is lower than that of Cl₂. The stoichiometric penalty is therefore controlled more by the CO₂ mass flow than by the chlorine mass flow. For a tower drawing 2,000 m³ h⁻¹ of ambient air at 20°C and 101.325 kPa with an average CO₂ concentration of 420 ppmv, the CO₂ volume flow is 0.84 m³ h⁻¹ and the CO₂ mass flow is approximately 1.54 kg h⁻¹. That CO₂ load alone requires 2.79 kg h⁻¹ of NaOH. If the same tower is absorbing only 5 kg h⁻¹ of chlorine, the chlorine-derived NaOH demand is 5.64 kg h⁻¹, so CO₂ increases the caustic feed requirement from 5.64 kg h⁻¹ to 8.43 kg h⁻¹, a 49% increase. This is why an atmospheric tower that is located outdoors and operated under slight vacuum should include an infrared CO₂ analyzer or a direct measurement of carbonate in the sump liquor using ASTM E291-18, rather than assuming that the chlorine feed-forward signal alone captures the full alkalinity demand.
The sodium carbonate formed by CO₂ absorption is not benign. In high-caustic environments, sodium carbonate may precipitate as the monohydrate or heptahydrate depending on temperature and ionic strength, fouling the lower packing, sump level bridles, and pH sensor diaphragms. The caustic feed ratio must therefore be high enough to cover the CO₂ reaction but not so high that the sump becomes supersaturated with carbonate. In a recirculating sodium hypochlorite system, a useful operational limit is to keep the total dissolved carbonate below 0.5 wt% as sodium carbonate; published data for this specific brine matrix is limited, but the precipitation tendency increases sharply when a 50 wt% caustic feed stream contacts a recirculating liquor already saturated with sodium chloride and sodium hypochlorite.
| Absorbed species | Reaction | Mass demand (kg NaOH per kg absorbed species) | Molar ratio (mol NaOH per mol species) | Operating consequence |
|---|---|---|---|---|
| Chlorine, Cl₂ | Cl₂ + 2 NaOH → NaOCl + NaCl + H₂O | 1.128 | 2.00 | Hypochlorite production; pH setpoint controls HOCl vapor loss |
| Carbon dioxide, CO₂ | CO₂ + 2 NaOH → Na₂CO₃ + H₂O | 1.818 | 2.00 | Carbonate scaling; caustic penalty from air ingress |
| Hydrogen chloride, HCl | HCl + NaOH → NaCl + H₂O | 1.097 | 1.00 | Immediate acid neutralization; localized pH drop |
In batch hypochlorite make-down systems for drinking water disinfection, caustic addition is frequently controlled by final pH rather than by inline flow ratio because the chlorine gas withdrawal rate from a small cylinder varies as the cylinder cools and as the liquid level in the batch tank changes. The operator target is usually a final pH of 12.0 to 12.5 with a free sodium hydroxide concentration of 0.3 wt% to 0.5 wt%, and the resulting product is analyzed for available chlorine and excess caustic according to AWWA B300-18 and ASTM D2022-89(2014). In such systems, the exothermic heat of chlorine absorption is removed by a plate heat exchanger using chilled water at 5°C to 10°C, and the batch temperature is held below 21°C. If the final pH drops below 10.5, the product decomposes rapidly and evolves chlorine; if the final pH exceeds 13.0, the product may attack seals and cause scaling with calcium and magnesium in the dilution water. The batch-to-batch variance is minimized by adding the caustic feed in a ratio-controlled stream until the pH reaches 11.0, then trimming with a metered caustic shot control to the final setpoint. This two-stage addition prevents overshoot and limits the time that the reaction mixture spends in the pH range between 9.0 and 10.0, where hypochlorous acid is a significant fraction and chlorate formation can accelerate.
The excess caustic inventory in the scrubber sump is not simply the difference between the caustic feed and the chlorine demand; it is the buffer that maintains the pH above the collapse threshold during a chlorine load transient. For a recirculating bleach tower with a sump volume of 5 m³ and an average liquor specific gravity of 1.20, a free sodium hydroxide concentration of 1.0 wt% corresponds to 60 kg of NaOH in the sump. If the chlorine feed is 12.7 kg h⁻¹, the theoretical NaOH consumption is 14.3 kg h⁻¹, meaning that the 60 kg buffer is exhausted in approximately 4.2 h if all caustic feed is lost. That time-to-collapse value is a useful design metric; many continuous towers are designed for a buffer time of 2 h to 6 h at maximum chlorine load, with the lower bound set by the response time of the feed-forward loop and the upper bound set by the cost of excess caustic and the risk of carbonate precipitation. The pH collapse threshold in a 15 wt% NaOCl solution is not a single universal value; it depends on temperature, chloride concentration, and total alkalinity. The control deadband around the pH setpoint is normally held at ±0.5 pH units, and the caustic feed pump is interlocked to trip on low sump pH at 10.0 or on high sump temperature at 27°C. This interlock is derived from the operating boundary that a low-pH excursion below 10.0 can produce chlorine desorption and a high-temperature excursion above 27°C can accelerate chlorate formation beyond the ability of the ratio loop to compensate.