Purging and Cleaning: Removing One Polymer From Another

What purging is actually removing

Purging is often described as cleaning, which sets the wrong expectation. You are not removing dirt. You are displacing one melt with another, and the difficulty is that the two melts do not want to mix. A purge is finished when the material leaving the die is the new material, at the right temperature, with no streaks, gels or unmelted fragments of the old one.

The volume that has to be displaced is larger than most operators expect. A 60 mm screw with a 24:1 barrel holds roughly 1.6 litres of melt. Add the adapter, the screen changer body, the melt pipe and the die, and the dead volume is often 2.5 to 3 litres. At a melt density around 0.9 g/cm3 that is about 2.5 kg of material sitting inside the machine at any moment, and all of it is old.

Displacement is not plug flow. The melt closest to the barrel wall moves slowly, and material in the flight clearance and in the corners of the adapter barely moves at all. Those slow regions are where the old polymer survives, and they are why a line can look clean for twenty minutes and then produce a streak from nowhere.

There is a second reservoir that costs more than the first: the material that cooled in the die during the last shutdown. It has to be melted out before anything can flow past it, and if the die was not purged before stopping, that is a separate job with a separate cost.

The three places polymer hides

The first is the screw channel itself, specifically the trailing edge of the flight and the clearance between flight and barrel. Material shears there continuously, so it is hot and mobile, but it is also the last to be exchanged. A screw with a worn flight, clearance above 0.4 mm, holds a film of old material that a normal purge will not shift.

The second is every change of section. Adapters, breaker plates, screen changer bores, melt pipes and die lips all have corners where the flow separates. A dead corner 3 mm deep and 40 mm long holds a few grams, and a few grams is enough to mark a thousand parts.

The third is the die land and the lip. This is the coldest part of the flow path and the one with the largest surface to volume ratio, so material hangs there and degrades. Degraded material from the die land is the usual source of the black specks that appear an hour after a colour change looks finished.

Knowing which of the three you are fighting changes the method. Screw channel contamination responds to a purge compound and to screw speed. Dead corners respond to dismantling and to a hot, low-viscosity flush. Die land responds to temperature and to physically wiping the lip while it is hot.

Why a purge compound works when resin does not

Running the new resin through until it looks clean is the oldest method and the most expensive in machine time. The reason is mechanical: two similar polymers with similar viscosities form a stable interface that travels down the screw without mixing. The interface stretches, folds and eventually breaks up, but slowly, and the tail of old material follows the new one for a long distance.

A purge compound attacks the problem differently. It is formulated with a carrier that is incompatible with most engineering polymers, plus a filler or an abrasive that scours the wall, and often a surfactant that wets the metal. It does not dissolve the old material; it lifts it off the surface and carries it as dispersed droplets.

The incompatibility is the point. A compound that mixes with the residue will dilute it instead of removing it, and dilution is exactly what plain resin does badly. This is also why the compound has to be matched to the family you are leaving, not the one you are going to: a compound designed for polyolefins will not lift a polyamide skin efficiently.

Temperature matters as much as chemistry. Purge compounds are usually run 10 to 20 C above the normal setpoint for the resin being removed, because lower viscosity means higher shear at the wall, and shear is what does the lifting. Running a purge cold to protect the material is a false economy that leaves the wall coated.

Matching the purge to the transition

Not every change needs the same effort. A light to dark colour change in the same polymer family is a displacement job that a small amount of compound will finish. A change from a glass-filled polyamide to an unfilled polypropylene is a different problem: the filler has to be flushed out of every corner, and any residual glass will show as a surface defect for hours.

The chart on the left is measured on a 60 mm single screw line changing from black glass-filled PA66 to natural PP. Running the new material straight through took 18 kg before the parts passed. A dedicated purge compound took 2.1 kg. Pulling the screw took 0.4 kg of material but three hours of labour.

The cost chart is the more useful one, because it includes what the material chart leaves out. Pulling the screw looks almost free on resin and costs EUR 364 once the fitter's time is counted. The purge compound costs EUR 115 all in. Running the material through costs EUR 749, and that number is why the habit survives: the resin is cheap and nobody books the machine time against it.

Table 1 — Which route suits which transition

Purging without stopping the line

On a line that runs the same product for weeks, the purge is a scheduled event and the cost is absorbed. On a line that changes material daily, the purge becomes a production constraint, and the way to control it is to make it part of the changeover rather than an interruption to it.

The single biggest saving is to purge before the machine cools. Material left in a hot barrel overnight degrades; material purged out at the end of the shift and replaced with a sealing resin comes back the next morning as a clean machine. The sealing resin should be one that is compatible with everything you run and cheap: PP or HDPE are the usual choices, and a low melt flow grade is better than a high one because it resists drooling.

The second saving is to purge with the die off. Removing the die before the purge cuts the dead volume, exposes the adapter for inspection and lets you see what is actually coming out. On a line where the die change is needed anyway, this costs nothing and shortens the purge by a third.

The third is to keep the screw turning during the flush. Stopping the screw to check the colour lets material sit and build a skin on the wall, and the next start pushes that skin into the flow as a flake. If you must stop, stop with the screw full and the heat on, and restart at low speed for thirty seconds before taking a sample.

Reading a bad purge from the parts

The first good part is not the end of the purge. A purge that has failed usually fails visibly within the first hundred parts, and the pattern tells you which of the three reservoirs is still dirty.

Streaks that run in the flow direction and fade over twenty or thirty parts come from the die land. The material there is being slowly replaced by the flow, and the fix is temperature and time, not more compound. Streaks that appear suddenly after a clean interval come from a dead corner that has just released, and the fix is to dismantle that corner.

Black specks that increase rather than decrease are the worst pattern, because they mean the barrel is degrading the new material. That is usually a sign that the purge ran too hot or that the previous material was PVC and left an acidic residue. It needs a mechanical clean, not another purge.

A drop in tensile strength or a change in impact resistance without any visible defect is the pattern that gets missed. It means a few percent of the old polymer is still blended in, and on a glass-filled transition that is enough to fail a specification. If the parts are going into a structural application, the purge is not finished until a mechanical test passes, not until the parts look right.

A purge procedure you can write down

A written procedure is worth more than a skilled operator here, because the variables are many and the feedback is slow. The version below is for a colour change within a polymer family and takes about forty minutes including the die change.

The order matters. Heat first, then purge, then check, then run. Heating after the purge compound is in the barrel means the compound sits and degrades, and a degraded purge compound leaves more contamination than it removed.

Table 2 — A colour change procedure for a single screw line

Record the resin used and the time taken for every change. After a few months the log shows which transitions are expensive and which compound suits them, and that is the information that makes the next change cheaper. It also shows when a purge that used to take 2 kg starts taking 4 kg, which is the first sign of screw wear rather than a change in the material.