Degassing & Refining Unit

What a Degassing & Refining Unit Is

A degassing and refining unit is an industrial melt-treatment system that removes dissolved hydrogen and non-metallic inclusions from molten light metal, primarily aluminium and its alloys, immediately before the metal is cast. It sits in the casting workflow after melting and holding, and before the filter box, launder, and casting machine. Functionally it is the metal-quality gatekeeper: everything downstream inherits the cleanliness it produces, and no later operation can re-dissolve the porosity it prevents.

The reason the unit exists comes from a single physical fact. Hydrogen is the only gas with meaningful solubility in liquid aluminium, and the amount that dissolves obeys Sieverts law, meaning the dissolved concentration is proportional to the square root of the hydrogen partial pressure above the melt. Water vapour is the main source: atmospheric humidity, damp charge material, wet refractory, oily scrap, and combustion products all react with aluminium to liberate atomic hydrogen. Hydrogen solubility in liquid aluminium can be described by the relation log S = 3.256 minus 2392 divided by absolute temperature, with S in cubic centimetres per kilogram, so solubility climbs steeply with superheat and roughly doubles for every 110 degrees Celsius of temperature rise.

The problem appears at solidification. Hydrogen is far more soluble in liquid aluminium than in the solid, by an order of magnitude near the freezing point, on the order of nineteen to one for pure aluminium. As the metal freezes, the rejected hydrogen has nowhere to go and precipitates as gas porosity, the rounded micro-voids that lower fatigue strength, cause pressure-tightness leaks, and ruin machined surfaces. Degassing before casting is the only practical way to keep dissolved hydrogen below the level at which this porosity forms.

Structurally, a degassing and refining unit has four functional parts: (1) a drive head with a variable-speed motor and a gas rotary union; (2) a graphite or ceramic shaft and rotor, the consumable element that is immersed in the melt; (3) an inert-gas supply with a mass-flow controller, and on fluxing machines a powder dispenser; (4) a control system that sequences rotor speed, gas flow, and treatment time, and on advanced units logs every cycle. When the rotor spins, it shears the incoming inert gas into a dense cloud of fine bubbles. Each bubble starts with essentially zero hydrogen partial pressure inside, so dissolved hydrogen diffuses across the bubble wall and is carried to the surface, where it burns off or vents.

The terminology can confuse buyers. Strictly, degassing removes dissolved gas, while refining is the wider task of removing solid inclusions and reactive trace elements. Because the spinning rotor accomplishes hydrogen removal and oxide flotation in one operation, and powder injection adds chemical refining, equipment is normally sold as a combined degassing and refining unit. Older names include rotary impeller degasser, rotary inert degassing or RID unit, and porous-plug or lance degasser for the simpler bubble-injection variants the rotary machines replaced.

Industrially, rotary degassing displaced two older methods: hexachloroethane tablets dropped into the melt, now banned in most regions for toxic emissions, and simple lance or porous-plug purging, which produced coarse bubbles with poor efficiency. The Spinning Nozzle Inert Flotation process, SNIF, developed by Union Carbide and now supplied through Pyrotek, and the Alpur in-line degasser from Novelis PAE established the in-line rotary standard for continuous casting, while compact box and ladle units such as the Foseco FDU brought the same physics to shaped-casting foundries.

Unit Types and Configurations

Degassing and refining units divide first by how they meet the metal: batch units treat a fixed quantity held in a vessel, while in-line units treat metal continuously as it flows past. The right family depends almost entirely on whether the foundry pours discrete batches or runs a continuous casting line. The table below compares the main configurations on the metrics that drive selection.

Ladle and mobile units are the simplest form. A wheeled or crane-mounted drive head lowers a rotor into a transfer ladle or crucible, runs a timed cycle of typically three to ten minutes, then withdraws. They are inexpensive, flexible across alloys, and ideal where metal is poured batch by batch, but each batch must be treated individually so throughput is limited by cycle time. Capacities commonly span a few hundred kilograms to about one tonne per batch.

In-furnace and box units, exemplified by the Foseco FDU family and equivalents from Pyrotek and StrikoWestofen, are permanently mounted over a holding furnace, dosing furnace, or dedicated treatment box. The FDU uses a patented rotor that creates fine inert-gas bubbles and disperses them homogeneously without disturbing the melt surface, and the process runs automatically once started, with control software setting rotor speed, gas flow, and treatment time per alloy and batch weight. These units suit shaped-casting and high-pressure die-casting melt shops that want repeatable, hands-off treatment of furnace-sized batches.

In-line units sit in the launder between the holding furnace and the casting machine and treat metal continuously as it flows through one or more sealed, refractory-lined chambers, each with its own rotor and inert-gas feed. The Spinning Nozzle Inert Flotation system, SNIF, and the Alpur degasser are the dominant licensed designs; both use multiple spinning nozzles or rotors, typically one to four, selected by required hydrogen reduction and throughput. The sealed chamber holds a protective inert atmosphere over the metal, which prevents re-gassing and lets the unit reach low hydrogen targets at high tonnage. In-line systems are the standard for direct-chill slab, billet, foil, and sheet casting where steady metal quality at high flow rate is essential.

Lance and porous-plug purging represents the legacy approach: inert gas is bubbled through a refractory lance or a porous plug in the furnace floor without a rotor. Bubbles are coarse, gas-to-metal contact area is small, and removal efficiency is poor, so these systems are increasingly replaced by rotary units except in low-specification remelt. They remain in service mainly for cost reasons and as supplementary purging.

A second axis of choice is automation. Entry-level units expose manual dials for speed, flow, and a timer. Mid-range units add programmable recipes per alloy. Premium units add process models, such as Foseco SMARTT, which compute the rotor speed, gas flow, and treatment time needed to hit a target hydrogen level given the batch weight, alloy, and starting condition, then run and log the cycle automatically. The same controller can perform deliberate upgassing with a nitrogen-hydrogen blend when a controlled hydrogen level is wanted to compensate for shrinkage in certain alloys.

Rotor Technology and Bubble Physics

The rotor is the heart of the unit and the single biggest driver of degassing efficiency. Rotary degassing works by maximising the surface area of inert gas exposed to the metal: for a fixed gas volume, smaller bubbles mean more total surface and faster hydrogen transfer. A good rotor shears the incoming gas into a fine, well-dispersed bubble cloud and distributes it through the whole melt volume without breaking the surface, which would otherwise entrain fresh oxide and re-introduce hydrogen. Rotor geometry, speed, immersion depth, and material all feed into this outcome.

Rotors fall into two material families with very different cost and life. The table below compares them on the properties that matter in service.

Graphite rotors use dense, fine-grain, high-purity graphite, typically with a bulk density around 1.70 to 1.86 g/cm3 and ash content below about 0.3 percent. Graphite is easy to machine into complex vane geometries, conducts heat well to resist thermal shock, and is inexpensive, which is why it dominates ladle and box units. Its weakness is oxidation: the shaft burns back where it crosses the metal line into air, and the vanes erode, so life is measured in tens to a few hundred cycles, often quoted as one to three months of single-shift use. Vacuum impregnation and protective coatings slow oxidation and extend service.

Silicon nitride rotors resist oxidation, thermal shock, and abrasion far better than graphite and can run two to five years, which transforms the consumable economics on a high-utilisation line. The trade-off is higher unit cost and brittleness, so they suit continuous in-line systems that run nearly around the clock rather than intermittent ladle work where thermal cycling and handling risk favour cheaper graphite. Silicon carbide occupies a middle ground, valued where abrasive inclusions or higher temperatures attack graphite quickly.

Beyond material, vane design is where manufacturers differentiate. Foseco markets patented XSR and FDR rotor profiles tuned to produce small bubbles and homogeneous dispersion, while the Alpur TS rotor is designed specifically to minimise bubble size and so increase the gas-melt interfacial area and the removal rate of hydrogen, alkalis, and inclusions. The practical lesson for a buyer is that rotor speed cannot be set by habit: the same rotor needs a different speed in a small crucible than in a large holding furnace, because the goal is fine dispersion without surface turbulence, and that balance shifts with vessel size and metal depth.

Two failure conditions bracket the operating window. Too little gas flow or too low a speed gives limited hydrogen transfer and slow degassing. Too much gas or too high a speed causes bubble coalescence into large slugs, surface turbulence, and oxide entrainment, which can leave the melt dirtier than it started. The correct setpoint is the one that produces the finest dispersed bubbles for the vessel, confirmed by a melt-quality test rather than guessed, which is why automated units that compute and log these parameters deliver more repeatable results than manual dials.

Inert Gas, Flux, and Process Media

The process media are the inert purge gas, optional solid fluxes, and the protective cover gas. Each choice affects degassing rate, alloy chemistry, and emissions. The dominant variable is the gas: it provides the bubbles that carry hydrogen out, and its purity, dryness, and chemistry determine whether the unit cleans the melt or contaminates it.

Nitrogen and argon are the two mainstream purge gases. Both are practically insoluble in aluminium and remove hydrogen by partial-pressure difference, so the choice is driven by cost and alloy compatibility rather than degassing physics. Nitrogen is much cheaper and is the cast-house default for most casting alloys. Its limitation is that at high rotor speeds, or with magnesium-rich and strontium-modified alloys, it can react to form aluminium nitride skins, so gas purity and dryness must stay high. Argon is fully inert under all conditions and is preferred for high-purity, grain-refined, and reactive alloys and for premium wrought product. Both gases should be at least 99.99 percent pure and dry, since any moisture in the supply line re-introduces the very hydrogen the unit is removing.

Reactive gases, historically chlorine or chlorine-nitrogen mixtures, were once added at a few percent to strip sodium, calcium, and lithium and to help break up oxides. They are now largely phased out in favour of solid powder fluxes because of corrosion, hazardous emissions, and worker-safety concerns, although small chlorine additions persist in some primary smelters under tight emission control.

Solid fluxes are powdered salt mixtures, commonly based on chlorides and fluorides of sodium, potassium, and aluminium, injected through the hollow shaft into the bubble stream or applied to the surface. Cleaning fluxes coalesce and detach oxides from the melt, cover fluxes form a protective blanket that reduces oxidation and hydrogen pickup, and drossing fluxes recover entrained metal from the dross. Injecting flux through the rotor combines mechanical bubble flotation with chemical refining in one cycle, which is the practical meaning of the refining half of the unit name.

The table below summarises the common media and where each is used. Always confirm specific flux chemistry and dosage against the supplier datasheet and the alloy being treated, since flux reactions are alloy-specific and overdosing wastes metal in the dross.

Refractory contact parts complete the media picture. The treatment chamber, baffles, and shaft sleeves are lined with non-wetting refractories chosen to resist aluminium attack and avoid shedding inclusions of their own, and seals on in-line chambers maintain the inert cover atmosphere that protects the treated metal until it leaves the unit.

Key Specification Parameters

A degassing and refining unit datasheet lists many figures, but a manageable set actually governs whether the unit will hit the metal quality you need. The parameters below are the ones to compare across quotations, grouped into process, mechanical, and verification metrics. The first table fixes the typical operating envelope; the discussion then decodes each.

Rotor speed typically runs from about 300 to 550 rpm and is the primary handle on bubble size. Higher speed shears finer bubbles up to the point where surface turbulence begins; the optimum is vessel-specific and should be commissioned with a melt-quality test, not copied between machines of different size.

Inert gas flow commonly falls in the range of roughly 10 to 60 litres per minute for ladle and box units, scaled up for large in-line chambers. Flow sets the total volume of bubbles. Process efficiency is often described by a gas removal ratio, for example a ratio of 200 means 200 litres of inert gas are needed to carry out one litre of hydrogen; the ratio worsens as the melt gets cleaner because the equilibrium hydrogen pressure inside each bubble falls, so the last increment of degassing always costs disproportionately more gas and time.

Treatment time for batch units is usually three to ten minutes and is the lever that turns a given gas flow and rotor speed into a final hydrogen number. Harder-to-degas alloys take longer: some high-magnesium casting alloys can take several times as long as pure aluminium to reach the same hydrogen level because their oxide skins and chemistry slow hydrogen transfer.

Hydrogen level is the headline result. Untreated foundry melts commonly sit at 0.20 to 0.40 mL of hydrogen per 100 g of aluminium. Correctly run units reach 0.10 to 0.15 mL/100 g for general castings, and below 0.10 mL/100 g for aerospace, wheel, and pressure-tight work. Single-pass removal efficiencies are reported around 40 to 70 percent depending on rotor design, gas, and cycle; multi-chamber in-line systems chain stages to reach the lowest targets.

Verification metrics belong on the spec because a hydrogen target is meaningless without a measurement method. Three are standard:

• Reduced Pressure Test (Straube-Pfeiffer): a sample solidifies under roughly 50 to 80 mbar absolute vacuum, expanding dissolved hydrogen so porosity becomes visible; weighing the sample in air and water gives a density index that ranks melt cleanliness quickly and cheaply on the shop floor.
• AlSCAN closed-loop recirculation analyser: recirculates a small inert-gas volume through a probe in the melt until hydrogen equilibrates, reading dissolved hydrogen in mL/100 g in minutes for in-line monitoring.
• ASTM E2792 inert-gas fusion: a laboratory referee method that fuses a solid sample and measures evolved hydrogen, used to calibrate and audit the faster shop-floor methods.

Mechanical and utility specs round out the comparison: drive motor power and speed control type, shaft and rotor material and expected life, gas mass-flow-controller range and accuracy, control level (manual, recipe, or model-based with data logging), fume extraction provisions, and the footprint and immersion depth that must match your vessel geometry. On in-line units, add chamber count, sealing and preheat arrangements, and metal-level control.

Selection Decision Factors

To turn the preceding chapters into a specific machine, work through the sequence below. Most selection errors come not from one wrong number but from deciding the wrong thing first, such as fixing on a rotor before settling whether the process is batch or continuous. Use these steps as a fixed RFQ template.

1. Batch or continuous, and throughput: First decide whether you pour discrete batches (favouring a ladle or box unit) or run a continuous casting line (favouring an in-line single or multi-chamber unit). Then fix the number: kilograms per batch and cycle frequency, or tonnes per hour matched to casting line speed. This one decision narrows the field more than any other.
2. Alloy family and gas: List the alloys to be treated and their hardest-to-degas member. Choose nitrogen for cost on standard alloys, argon for high-purity, reactive, or grain-refined alloys, and verify gas compatibility with any magnesium-rich or strontium-modified grades that risk nitride formation.
3. Hydrogen target and verification: Set the acceptance hydrogen level from the casting specification (general, automotive, or aerospace) and decide the measurement method up front: Reduced Pressure Test and density index for shop-floor control, AlSCAN for in-line monitoring, and ASTM E2792 for referee audits. The target plus the starting hydrogen defines the removal duty.
4. Rotor material and consumable economics: Match rotor material to utilisation. Graphite for intermittent ladle and box duty where it is cheap to replace; silicon nitride or silicon carbide for high-utilisation in-line lines where long life offsets the higher price. Cost the rotors and shafts over a year, not per piece.
5. Refining and fluxing needs: Decide whether alkali removal and inclusion fluxing are required, which calls for a unit with through-shaft powder injection, or whether degassing and bubble flotation alone suffice. Confirm flux chemistry and dosing with the supplier for each alloy.
6. Automation and data: Choose the control level. Manual dials for simple low-spec work; recipe-based control for repeatability across alloys; model-based control with logging (for example Foseco SMARTT) where traceable, target-driven treatment and process records are required for quality systems.
7. Integration with the melt-treatment train: Confirm how the unit fits with melting, holding, ceramic foam filtration, and the casting machine. A degasser does not replace a filter: fine inclusions below roughly 20 micrometres still need a 30 to 80 ppi ceramic foam filter downstream, so specify the whole train together.
8. Safety, fume, and utilities: Verify fume extraction for flux and burn-off gases, inert-gas supply pressure and purity, electrical and cooling utilities, and operator protection around hot rotors and molten metal. These determine installation cost and compliance as much as the machine price.

One last dimension that buyers underweight is serviceability and consumable supply: lead time and local stock for rotors and shafts, refractory chamber relining intervals on in-line units, availability of trained commissioning to set rotor speed and gas flow for your vessels, and spare-parts support over a ten-year service life. Established suppliers such as Foseco, Pyrotek, Novelis PAE (Alpur), StrikoWestofen, and ABB (AlSCAN measurement) maintain consumable and service networks; a low purchase price means little if a worn rotor takes weeks to replace and the line stops.