Regenerating an acetylene purifier: the 72-hour air cycle the 45 °C stop and when a bed is finished

A dry acetylene purifier is a chemical bed, and like any chemical bed it is consumed. The ferric chloride mass that strips phosphine and hydrogen sulphide from the gas is re-oxidised with air, roughly ten times over its life, and each cycle restores a little less. Knowing how to run the 72-hour regeneration, when to trigger it and when to stop pretending the bed can be revived is most of what separates a plant that fills clean cylinders from one that does not.

Gas from a carbide-to-water generator is already 99 to 99.5 % acetylene on a dry basis; the main contaminant is air entrained with the feedwater. Purification in the trade sense means removing the non-metallic hydrides, mainly phosphine and hydrogen sulphide, with arsine and ammonia alongside. Phosphine comes from calcium phosphate in the lime and coke reduced to phosphide in the carbide furnace; hydrogen sulphide from calcium sulphide. The lime slurry in a wet generator holds back most of the sulphur, so hydrogen sulphide in crude gas is usually below 10 vpm, but phosphine passes almost entirely into the gas.

For dissolved acetylene IS 308:2024 sets 0.01 % by volume for phosphorus compounds as PH₃ and 0.02 % for sulphur compounds as H₂S. A ferric chloride bed of the Regenetal type can bring phosphine down to about 0.1 vpm and hydrogen sulphide to about 1 vpm at a working space velocity of about 50 bed volumes per hour, as compiled by S. A. Miller (1965) from British practice. That margin is why the silver nitrate paper, which responds to 5 vpm instantly, is such a good breakthrough detector.

The real reason to purify is the cylinder, not the weld. Phosphine and divinyl sulphide are retained by the porous mass and acetone; of the phosphorus charged with gas at about 0.06 % PH₃, only 55 to 79 % came back out over 14 days depending on temperature (Hölemann and Hasselmann, Forschungsberichte des Landes Nordrhein-Westfalen No. 739, 1959). What stays accumulates for the life of the cylinder.

A British Oxygen Company survey, run by the BS 642:1951 method on carbide offered for sale in Britain over ten years and reported by Miller, found phosphine from 19 manufacturers in 12 countries all within 150 to 310 vpm, with Indian carbide at 800 to 1,000 vpm, the highest in the table. The figures date from the 1950s and 1960s and say nothing about any present supplier, but they explain a design fact: an Indian purifier bed does three to six times the phosphine duty of a European one per cubic metre of gas. The bed is also the plant's effective lime-dust filter, taking entrained lime from about 21 mg/m³ down to about 1 mg/m³ in the first hour after start-up (Hölemann, Forschungsberichte des Landes Nordrhein-Westfalen No. 888). Stop-start plants load it faster still. Both facts argue against a calendar interval and for a breakthrough-triggered one.

KIGL purifiers of the non-regeneration type take a charge of 250 kg on the DA-25 and 400 kg on the DA-45. The material is sieved through an 8 to 10 mesh screen and laid in three layers, each on its own stainless steel wire mesh: a first layer of about 400 mm, a second of 400 mm, and a top layer of about 200 mm of lime-kieselguhr mixture. The mass is spread with a light wooden rake and the edges pressed with a wooden reamer; the vessel interior is protected with zinc chromate primer and two coats of anti-acid paint.

The top layer deserves a word, because operators routinely misunderstand it. The lime is an HCl guard bed. Ferric chloride masses originally introduced traces of dichloroethylene and hydrogen chloride into the gas; the dichloroethylene was eliminated by cutting the mercury and copper chloride content to the minimum needed, and a layer of lime on top of the bed holds back any trace of HCl. It is not a filter, not a drier, and not an active purifying layer.

Before acetylene is admitted to a fresh or regenerated bed the vessel is purged with dry nitrogen for 2 to 3 minutes and the cover joint soap-tested. The purifier then runs at 0.1 to 1.3 kg/cm²g and within about 15 °C of ambient temperature, with its bottom drained every 8 hours and the drain valve never left open.

The rotation is run with progressively shortened intervals, so the record of each purifier's service dates is what tells the plant when repacking is due. The records the manual requires are not optional: date of repacking and quantity used, date out of service, date returned, hours on the fan and maximum air temperature, every purity test result, and any fault. They are open to inspection at any time, and they are the only evidence a plant has that its gas met IS 308 on a given day.

1. Isolate the purifier by closing its gas inlet and outlet valves slowly, then purge it with nitrogen so that no acetylene-air mixture can form inside.
2. Connect the regeneration blower. The fan must deliver at least 15 cm WG and the blower at least 18 cm WG; the air pressure in the vessel is kept low, within the limit set in the plant's own operating manual, because a purifier tank blown hard enough can get hot enough to start a fire.
3. Run the air for 72 hours.
4. Record the outlet air temperature every hour for the first 4 hours and every 2 hours thereafter until it stops rising.
5. If the outlet air reaches 45 °C, stop the fan at once and let the bed cool for one hour, or cool it with nitrogen at less than 0.4 kg/cm²g, before restarting the air.
6. For the first 4 hours after the bed returns to service, feel the vessel by hand for warm spots; if any are found, keep nitrogen on at less than 0.4 kg/cm²g until cool, then open the vessel and replace any discoloured material.
7. Drain the regeneration pipeline near the exhaust fan at the same 8-hour interval as the purifier bottoms.

The 72 hours is not a KIGL invention. British dissolved-acetylene practice regenerated ferric chloride masses in place with an air current at a space velocity of about 11 bed volumes per hour for 72 hours, against about 50 per hour in gas service. The air is deliberately four to five times slower than the gas because re-oxidising ferrous to ferric iron is exothermic; the 45 °C stop and the low air-pressure limit are the same control expressed as a temperature and a pressure rather than a flow. A blower sized on the gas rate will run the bed hot.

Air restores the oxidant. It does not remove the products. The phosphate, phosphite and sulphate formed in service stay on the kieselguhr, progressively occupying surface and blocking pores, and the lime dust the bed has filtered out of the gas stays with them. That is why a ferric chloride mass can be regenerated about ten times before it is cheaper to discard, with a somewhat shortened run after each regeneration. The two signs that a bed is approaching the end are a purity test that fails sooner each cycle and a pressure drop that no longer recovers.

Spent mass is a mercury-bearing, acidic solid: mercury as chloride, phosphide and sulphide, iron and copper chlorides, calcium phosphate and sulphate, and the HCl-laden lime from the top layer. It is removed with the lid and lining treated carefully. The manual's sequence is to purge the acetylene out with nitrogen (older KIGL manuals say air or nitrogen; nitrogen is the only purge gas for acetylene plant), run the fan for not less than 12 hours, blank all regeneration lines, mark the lid-to-tank alignment, hoist the lid, remove the top screen one half at a time and never scratch the lining. Rubber gloves, and a dust mask if the mass is dry, are mandatory. The disposal route is a hazardous-waste question that the plant's own licence conditions must answer.