Caustic Scrubber Efficiency Loss After a Packing Retrofit
A packed caustic scrubber treats 8,000 m³/h of vent gas at 25 °C and 1 atm containing 500 mg/Nm³ HCl. The 1.2 m ID column has 4 m of 50 mm plastic random packing irrigated with 5 % NaOH recirculated at 12 m³/h; the sump is held at pH 12.5. The original packing was 38 mm and was replaced six months ago with 50 mm to cut pressure drop from 1.2 to 0.6 kPa/m. Since then stack HCl has risen from 18 to 145 mg/Nm³ against a 30 mg/Nm³ limit, and caustic consumption has doubled. Sump pH stays high and no flooding is observed. The liquid distributor, which has 28 drip holes, was not changed. Spent caustic goes to the biological effluent plant at a cost of €0.9/kg NaOH.
1. Diagnose the cause of the efficiency loss and explain why raising the NaOH concentration or the sump pH setpoint will not fix it.
Read the worked answer
Required versus achieved removal: required = (500 − 30)/500 = 94 %; achieved = (500 − 145)/500 = 71 %. This is a large loss, not a marginal drift.
Why pH cannot be the fix. HCl absorption into NaOH is an instantaneous reaction in the liquid film. The rate is gas-film controlled: N = kG·a·p(HCl). Once the liquid is alkaline (pH well above 8–9), the interfacial HCl concentration is effectively zero and the driving force is set by the gas side, not by bulk pH. Raising the sump from 12.5 to 13 therefore changes almost nothing; it only increases the caustic sent to drain. The observation that pH stays high while consumption doubles confirms that caustic is being wasted elsewhere (carbonate blowdown, CO2 absorption), not being used to capture HCl.
Why the retrofit caused it. Two effects compound:
- 50 mm random packing has roughly 20–30 % less geometric surface per m³ than 38 mm, so the volumetric mass-transfer coefficient kG·a falls.
- The unchanged 28-hole distributor over 1.13 m² gives only ~25 drip points/m², below good practice for this packing size, and the irrigation density is 12/1.13 ≈ 10.6 m³/(m²·h), at the bottom of the acceptable band. Poor initial distribution causes channelling and dry patches, so the effective wetted area falls much more than the geometric area.
The low pressure drop is a symptom of a column that is under-loaded and under-wetted, not proof of good hydraulics. Before spending money, confirm with a gas-phase HCl profile up the bed and a distributor inspection; look for a dry or crusted zone and for caustic mist carryover, which can also raise the measured stack number.
2. Do the caustic balance and decide between restoring 38 mm packing, increasing the liquid rate, or adding a second stage.
Read the worked answer
Stoichiometry:
- HCl load = 8,000 × 0.5 g = 4.0 kg/h = 0.110 kmol/h.
- NaOH required = 0.110 kmol/h × 40 = 4.4 kg/h stoichiometric; at 20 % excess, 5.3 kg/h.
- Stack loss at 145 mg/Nm³ = 1.16 kg/h HCl, i.e. about 29 % of the load escaping - consistent with the measured efficiency.
Why consumption doubled. CO2 in the vent at even 400 ppm is 3.2 m³/h = 0.143 kmol/h, which consumes 0.286 kmol/h of NaOH = 11.4 kg/h, more than twice the HCl demand. Caustic consumption is therefore dominated by CO2 and by blowdown to control carbonate, not by HCl. Chasing the consumption number will mislead the investigation.
Options:
- Increase liquid rate. kG·a rises weakly with liquid rate and the column is already near the low end of irrigation density; a modest increase (to 18–20 m³/h) may recover a few points but is unlikely to bridge 71 % → 94 %, and it risks flooding if the distributor is poor.
- Restore high-surface packing. This addresses the main cause but keeps the defective distributor, so the gain will be partial.
- Second stage. Effective but capital-heavy and treats the symptom.
Recommendation: replace the distributor with a trough or pan design giving at least 60–80 drip points/m² and a documented turndown, then reinstate high-efficiency packing (38 mm random or a structured packing with a validated kG·a). Keep the pressure drop low by sizing the packing for the actual F-factor rather than by coarsening it. Verify with a performance test at design and 50 % load, and add a mist eliminator check.
3. Write the operating and environmental controls you would put around the scrubber after the fix.
Read the worked answer
Control:
- Feedforward from vent-gas flow and an inlet HCl analyser to the caustic makeup, with pH as a slow trim. A pH loop alone has a logarithmic response and a long dead time and will overshoot the stoichiometric point.
- Recirculation flow indication with a low-flow alarm interlocked to the vent-gas diversion, because a dry bed gives almost no removal.
- Continuous or daily stack HCl monitoring with a documented detection limit well below 30 mg/Nm³; a colorimetric tube is not adequate for a permit limit.
- Conductivity or density-based blowdown control to hold carbonate at a set level instead of blind periodic dumping.
Environmental:
- The spent caustic is a high-TDS, high-pH stream. Confirm the biological plant's consent limits for sodium, chloride and pH, and the hydraulic capacity; a doubling of blowdown may breach the trade-effluent consent even though the stack is compliant.
- Caustic mist carryover is both an emission (particulate, visible plume) and a corrosion risk downstream. Check the mist eliminator and the stack drain chemistry.
- Record the caustic consumption per tonne of product as a key performance indicator; a rise is the earliest signal of distributor or packing degradation.
Change control: the packing and distributor change altered the scrubber's mass-transfer performance and the permitted emission point. It should have been an MOC with a performance guarantee and an acceptance test. Add that test to the standard for any future packing change.