Pick-Up and Die Lines in Aluminum Extrusion: Three Defects, Three Causes
Three surface defects account for most of the scrap on a 6xxx profile: pick-up, die lines and micro-die lines. They look alike on the extrudate and they do not have the same cause. Pick-up is aluminum building on the exit side of the die land and breaking away onto the profile. A die line is stamped into the surface at the front of the land, where the extrudate still chokes the throat. A micro-die line is not made by the die at all. It is a string of voids that opens in the billet’s own surface layer. Telling them apart decides who fixes the problem, and whether the fix is a die, a billet or a press setting.
The three defects, and where each one starts
| Defect | What it looks like | Where it comes from | Who can fix it |
|---|---|---|---|
| Pick-up | Intermittent score lines, 3 to 12 mm long, usually ending in a fleck of aluminum that has been ironed into the surface | Aluminum coating on the exit half of the die land, breaking away and transferring to the extrudate | Die shop (land and nitriding), press settings, billet cleanliness |
| Die line | Continuous longitudinal groove or ridge, deep enough to survive etching, sometimes visible before anodizing | The choked front of the land, roughly the first 2 mm, where the surface is generated against the aluminum film | Die shop (land length, land finish, choke, die deflection) |
| Micro-die line | Fine, shallow, continuous line. Survives polishing. Usually first seen after anodizing | Fractured iron-bearing precipitates and the voids between them, in the surface layer of the extrudate | Billet and homogenization, not the die |
Pick-up is the defect an operator sees first, because the fleck sits proud of the surface and can be felt. A die line is easier to miss at the press and much harder to hide afterwards. A micro-die line is the one that returns a load of anodized profiles, because the etching that precedes anodizing opens it up.
The surface is made in the first two millimetres of the land
The land does not work as one surface. At the entry the extrudate is pressed hard against the die and sealed from the atmosphere, and the aluminum skin that forms there is the surface the customer receives. Behind that point the extrudate lifts clear as the die deflects under load, air reaches it and an oxide forms, and the rest of the land does nothing but iron, gall and score the surface it has already made.
The trials behind this used lands ground perpendicular to the extrusion direction at 1.1 µm Ra, from 8 mm down to a knife edge. The aluminum build-up on the working area of the land measured about 2 mm on every die in the series, whatever the total land length. Two results follow. A land under about 1 mm is not stable under extrusion load and chatters, and the surface comes out rougher than it does with 2 mm, so a knife edge is not a free finish. And every millimetre of land beyond the choke adds friction, heat and transfer area without improving the surface. About 2 mm, fully choked, gave the best finish in the series.
That is also why the land finish matters more than the land length in most shops. At the optimum temperature the aluminum film sits down inside the grinding valleys instead of standing above them, so the land behaves as a smooth surface and the product follows the film rather than the tool marks. At 550 °C the same film stands proud of those valleys and the product takes its ridges. The land’s roughness profile is part of the die, not a finish applied to it.
450 °C makes the surface, 550 °C breaks it
What the film on the land looks like depends on temperature, and the film is what decides the finish. The same 6063 billet, run at the same exit speed, produced three different surfaces across a 200 °C span.
| Billet temperature | What the aluminum film does on the land | Result on the profile |
|---|---|---|
| 350 °C | Mg2Si is rejected as coarse precipitates, so the small hard particles that scour the land are gone. Flow stress is higher, extrusion pressure is higher, and the die deflects more, which lets more aluminum build | Roughness rises and the scatter widens. Coarse bands of aluminum in the choke print onto the surface |
| 425 to 475 °C | The film sits inside the grinding valleys of the land instead of standing above them. The land behaves as a smooth surface, and fine (Fe,Mn)Al3 and Mg2Si particles scour it clean | Smoothest surface of the series, about 0.5 µm Ra as extruded. This is the industrial range for 6063 |
| 550 °C | The film stands proud of the ground surface as ridges. It work-hardens and is quenched by the die, so it becomes harder than the extrudate passing over it | The profile takes the shape of the ridges, and heavy pick-up collects at the land exit |
Surface finish against billet temperature for 6063, at an exit speed of 1 m/s. Three billets were run at each temperature with the lands cleaned before the first. The as-extruded finish at the optimum was approximately 0.5 µm Ra.
The useful part is the shape of the curve, not the absolute number. There is an optimum in the middle, and both ends fail for their own reason: too cold, and the particles that would scour the land are locked up as coarse precipitates while the die deflects further; too hot, and the film grows into ridges that are harder than the product. Moving a die from 500 °C to 575 °C makes pick-up worse, not better, because the film elongates, spalls on both sides and then scores the extrudate until it too transfers.
Pick-up: two mechanisms, three levers
Pick-up forms in one of two ways. In the first, adhesion between the die and the extrudate is roughly equal to the cohesive strength of the aluminum at the surface, so flakes transfer back and forth between land and profile until they agglomerate into a deposit. That is the stick-slip mechanism, and it is common when a new die is commissioned. In the second, aluminum collects at the exit of the land, work-hardens as it is dragged, elongates in the extrusion direction and finally spalls, leaving a deep score with the deposit at its end. In 6063 the second is the usual one.
Severity follows temperature and strain rate together, so a temperature on its own does not define a safe operating point. The published form is a line in the temperature–strain rate plane: the surface stays acceptable while the temperature-compensated strain rate Z stays below ln Z = aT − b. The constants depend on alloy, cast condition, homogenization, die design and die temperature, which is why a threshold copied from another plant rarely transfers. The form does.
Three levers sit on the tooling side. Nitriding the land changes the surface the aluminum sees. A non-wetting die material, such as a sialon insert, removes the adhesion itself. Die cooling pulls the land temperature down without touching the billet. Two more sit upstream: a homogenization treatment that spheroidizes the iron phase, and a billet in which no coarse Mg2Si is left in the surface layer. Incomplete spheroidization after a short homogenization is enough to make pick-up worse under conditions that would otherwise pass.
Micro-die lines are a billet defect, not a die defect
A polished die, run at the optimum temperature, still produces die lines. They are finer and shallower than the ones from a rough land, and the industry calls them micro-die lines, but the name hides the fact that the die is not producing them.
The sequence runs in the billet’s surface layer. Coarse iron-bearing precipitates align with the flow as the material enters the shear zone outside the dead metal zone. Near the die throat they are sheared and fractured, voids open between the fragments and link into a string, and when that string reaches the surface it reads as a fine score line. The same surface layer also picks up small Mg2Si particles released where the slow-moving material on the die face meets the faster flow behind it, which is why the effect is concentrated in the first few tenths of a millimetre below the surface and nowhere else in the section.
None of this responds to die steel. Polishing cannot remove a void below the surface, nitriding cannot close it, and a cleaner grade or a coating does not touch it. The levers are homogenization, billet surface quality and die design at the face, and this is the one defect on the page where a die shop should hand the problem back to the billet supplier.
What appears after anodizing
Anodizing does not create these defects, it reveals them. The etch removes a controlled amount of metal, and any difference in microstructure between one band of the surface and the next becomes a difference in tone. Most of the complaints that reach a die shop after anodizing started somewhere else.
| What the customer sees | Origin | What to check |
|---|---|---|
| Streak following the extrusion direction, usually on the wide face | Ingot shell zone or surface segregation. The shell zone can run up to 15 mm deep, and the faster flow into wide parts of the section feeds it to the surface | Billet source and casting practice, scalping, die layout against the container diameter |
| Streak that matches a die feature, sometimes with a grain size difference | Die design. Local differences in strain rate change the temperature and the microstructure in that band | Bearing layout and ports, not the billet. Mechanical treatment of the profile will not remove it |
| Streak along a weld line | Weld in a hollow or feeder-plate die. The microstructure change is larger than from die design alone | Weld chamber and port balance |
| Streak or dark band near the back end of the length | The ram came too close and the billet skin became the core of the extrusion | Discard thickness against billet diameter |
| Dark, rough patches at intervals, softer than the rest | Local chilling on the run-out table, then reheat into the Mg2Si precipitation range at about 400 °C | Contact with carbon support blocks, cooling rate after the press. Not recoverable once formed |
| Grainy, galvanized look | Zinc in the alloy or in the etch bath. Noticeable above 0.03 % Zn and worse towards 0.1 % | Metal source, etch bath chemistry and temperature |
What the die shop can fix, and what it cannot
Inside the die shop: land length and choke, land finish and polishing direction, nitriding, die temperature and die cooling, bearing relief, and the support behind the die that decides how far it deflects. Outside it: micro-die lines, shell-zone streaks, spangle, hot spots from the run-out table, and pick-up driven by a cold billet or a coarse Mg2Si surface layer.
The test that separates them is cheap. A defect that survives a freshly polished and freshly nitrided die, run inside the 425 to 475 °C window, is not a die problem. Spending on a cleaner or harder grade at that point changes nothing on the profile.
How a die is corrected, and what to record when it is, is on the extrusion die correction page. The press window that sets the bearing temperature is on the extrusion temperature page.
What the billet contributes to those defects, and which billet numbers a foundry can hold, is on the billet quality page.
Where this leaves a die steel order
Pick-up is an adhesion problem on the bearing, and the bearing is the one place on the press where the steel is doing the work. That is why the bearing runs at 46 to 50 HRC and why it is nitrided to a case of 0.10 to 0.20 mm on a chromium hot work steel rather than a cold work grade. The 1.2344 ESR programme is on the 1.2344 ESR H13 extrusion supply page, the grade itself on the H13 tool steel page, and the case on the tool steel nitriding page.
What the grade does not do is worth stating plainly. It does not replace homogenization, it does not rescue a land that is too short, and it does not remove a micro-die line. Where the defect has already been traced to the billet, a different steel only moves the cost. The bearing that wears is on the die bearing length page, the mechanisms that consume a die on the die wear page, the stack around the bearing on the die types and tooling page, and the grade by component on the extrusion die steel selection page.
Before you use this as a surface specification
The figures on this page come from controlled trials on 6063 through split-die tooling, not from a production press. The optimum temperature, the minimum land length and the roughness bands all shift with alloy, section, die temperature and press, and the constants in the pick-up relation are specific to the alloy and the cast condition. The values are here to show which variable moves the surface and in which direction. Final surface condition is confirmed on the job.
Source, Extrusion of Aluminium Alloys, T. Sheppard, Kluwer Academic Publishers, 1999, sections 6.4 and 6.7.