Metal Extrusion

Metal extrusion in one picture

Picture a hydraulic press the size of a shipping container. At one end sits a steel die — a thick disc with a hole cut to the exact cross-section you want, say the silhouette of a window frame. Behind the die is a heavy-walled container, and behind that a ram. You drop in a glowing aluminum billet — a solid cylinder heated to about 480 °C — and the ram drives forward with thirty meganewtons of force. The metal, soft and plastic but nowhere near molten, has nowhere to go but through the die hole. It flows out the front as a continuous length of profile that is an exact copy of the die opening, growing at forty meters a minute until the billet is used up.

That is the whole idea: shape the hole, push metal through it, and the metal takes the shape of the hole. It is to solid metal what a pasta machine is to dough, or what squeezing toothpaste is to the tube. The genius of the process is that all of the shape-making lives in one cheap, reusable tool. A die that costs a few thousand dollars and weighs a few kilograms can run tens of thousands of meters of identical profile before it wears out, which is why extrusion is the default way to make any long part with a complicated but constant cross-section.

The constraint is exactly that: the cross-section must be constant along the length. Extrusion cannot make a part that tapers or changes shape from one end to the other — that is the job of forging, casting, or machining. But for the vast universe of parts that are "the same shape, just long" — framing, rails, tubes, bus bars, finned heat sinks, mullions, structural channels — nothing competes with it on cost per meter.

The mechanics — strain, force, and the log law

When the ram drives the billet through the die, every bit of metal is squeezed from the billet's large area down to the profile's small area. The amount of squeezing is captured by the extrusion ratio:

R = A_billet / A_profile

and the true (logarithmic) strain the metal undergoes is simply:

ε = ln R

The ram pressure needed to push the metal through follows directly. To a first approximation, ignoring friction and redundant work, the pressure is the flow stress times the strain:

p = σ_flow · ln R          (ideal, frictionless)
p = σ_flow · (a + b·ln R)  (Johnson empirical, a≈0.8, b≈1.2–1.5)

The Johnson constants a and b bundle together friction and "redundant work" — the energy wasted shearing the metal internally as it bends through the die rather than flowing in a clean straight line. The key consequence of the logarithm is that doubling the extrusion ratio does not double the pressure; going from 10:1 to 20:1 only raises ln R from 2.30 to 3.00, a 30 percent increase. That is why very high reductions are practical in one pass.

Let's run the numbers for a real architectural profile. Take a 6063-T5 aluminum billet 178 mm in diameter (A_billet ≈ 24,900 mm²) extruded to a window-frame section of total area 500 mm², so R ≈ 50:1 and ε = ln 50 = 3.9. At 480 °C the flow stress of 6063 is roughly 40 MPa. Using the Johnson form:

p = 40 MPa × (0.8 + 1.3 × 3.9) = 40 × 5.87 ≈ 235 MPa
Ram force F = p × A_billet = 235 MPa × 24,900 mm²
            = 5.85 MN  →  with friction/redundant work, ~30 MN class press

The bare metallurgical pressure comes out modest, but real presses are sized two to five times higher to cover billet-container friction, die-bearing friction, breakthrough peaks at the start of the stroke, and a safety margin. A 30 MN ("3,000 ton") press is a typical mid-size commercial extruder; the largest forward-extrusion presses in the world reach 200 MN for big aerospace and rail sections.

Direct vs indirect — where the friction lives

There are two fundamental geometries, distinguished entirely by whether the billet slides along the container wall.

• Direct (forward) extrusion. The die is fixed at the front; the ram pushes the billet forward through it. The entire billet slides along the container, so billet-container friction is huge. The ram force is highest at breakthrough (longest billet, most wall contact) and falls as the billet shortens — the classic falling force-vs-stroke curve. Simplest, fastest, and overwhelmingly the most common.
• Indirect (backward) extrusion. The die rides on the end of a hollow stem and is pushed into a closed container; the billet never moves relative to the wall. With no billet-container friction the force is 25–30 percent lower and roughly flat across the stroke. The metal also runs cooler and gives more uniform end-to-end properties. The penalty is the hollow stem, which limits profile size and complicates handling. Reserved for hard alloys (7075), copper, and jobs needing tight property uniformity.

A third laboratory and specialty variant is hydrostatic extrusion, where the billet is surrounded by a high-pressure fluid that both pushes it and lubricates the die, eliminating container friction entirely and allowing brittle materials to be extruded that would crack under ram contact. It is slow and expensive, used for things like superconducting wire and beryllium.

Hollow profiles — the porthole-die welding trick

A solid die just has a hole. But how do you extrude a hollow tube or a multi-void window frame, where there has to be metal on the inside and the outside, with empty space between? The answer is one of the most elegant tricks in manufacturing: the porthole (bridge) die.

The die carries a central mandrel held by structural "bridges". As the billet is forced in, the metal is split into two to four separate streams that flow around the bridges through openings called ports. Immediately downstream the streams are forced back together in a welding chamber under colossal pressure — often above 500 MPa. Because the metal is hot, the freshly created internal surfaces have no oxide, and the pressure is enormous, the streams solid-state weld back into a seamless wall as they pass over the mandrel. The mandrel forms the inside surface; the surrounding die ring forms the outside. One tool, one pass, a closed hollow section.

The weld lines (called charge welds or seam welds) are the structural weak points, and they are why this trick only works for the readily weldable alloys — the 6000 series above all. The high-strength 7000-series alloys and copper do not seam-weld reliably, so their tubes are made over a separate piercing mandrel or by indirect tube extrusion. Every aluminum window frame, every rectangular structural tube, every multi-channel heat-sink-with-internal-fins you have ever seen was almost certainly made through a porthole die.

Why aluminum owns extrusion

Roughly 80 percent of all extruded metal tonnage is aluminum, and the reasons are physical. At its 450–500 °C working temperature aluminum's flow stress is only 30–80 MPa, it does not gall or stick aggressively to steel tooling, and the workhorse 6000-series alloys (6061, 6063) are both genuinely strong and extremely extrudable. A 6063 profile runs through the die at 40–60 m/min. Crucially, aluminum age-hardens: the profile is extruded hot, quenched right at the press exit (press quenching), and then artificially aged in an oven to the T5 or T6 temper, recovering its strength so it leaves the line ready to use.

Metal / alloy	Billet temp	Relative force	Typical speed	Notes
Aluminum 6063	450–500 °C	1× (baseline)	40–60 m/min	Most extrudable; window frames, heat sinks
Aluminum 7075	400–450 °C	2–3×	1–5 m/min	Hard; often indirect; aerospace sections
Copper / brass	650–950 °C	3–5×	10–30 m/min	Bus bars, tube; no porthole welding
Steel	1100–1300 °C	5–10×	2–10 m/min	Needs glass lubricant (Ugine-Sejournet)
Titanium	900–1200 °C	6–12×	1–5 m/min	Niche; aerospace; glass lube, fast quench
Magnesium	300–400 °C	1–2×	5–30 m/min	Light sections; flammability care

Defects — how extrusions go wrong

• Extrusion (pipe) defect. In direct extrusion the hot core flows faster than the cooler, oxidized billet skin. Near the end of the stroke the skin gets sucked into the center, leaving a funnel-shaped internal void running down the back of the extrudate. Cure: stop short and discard the last 10–15 percent as the butt.
• Surface (speed) cracking — "fir-tree". Pushing too fast or too hot drives the exit temperature past the alloy's incipient melting point; the surface tears in a repeating chevron pattern that looks like a row of fir trees. Cure: slow down, pre-cool the billet, or run isothermal extrusion that ramps speed down as the deforming metal heats up, holding exit temperature constant.
• Center-burst (chevron) cracking. Too low an extrusion ratio with a poor die geometry produces tensile hydrostatic stress on the centerline that opens internal voids — invisible from outside, catastrophic in a structural part. Cure: adequate ratio, good die-land design, and back pressure.
• Charge-weld weakness. In porthole hollows the seam welds are the weakest line; insufficient welding-chamber pressure or oxidized surfaces give weak, sometimes leaking, welds. Cure: more pressure, clean billets, weldable alloy.
• Die-line streaking and pickup. Tiny scratches on the die bearing drag along the whole length, leaving visible longitudinal lines; aluminum can also build up ("pickup") on the bearing and tear the surface. Cure: nitrided dies, polished bearings, and regular die maintenance.

Extrusion vs the alternatives

Property	Metal extrusion	Rolling	Forging	Casting
Output shape	Constant complex cross-section, long	Flat / simple section, long	3D net shape, discrete	Any 3D shape, discrete
Cross-section complexity	Very high (hollows, fins)	Low	Medium–high	Very high (incl. internal)
Tooling cost	Low (cheap dies)	High (roll stands)	High (die sets)	Medium (molds)
Material strength	High (wrought grain along length)	High (wrought)	Highest (grain flow follows shape)	Lower (cast porosity)
Length limit	Effectively continuous (per billet)	Continuous (coil)	Short parts	Mold-bound
Best for	Frames, heat sinks, rails, tube	Sheet, plate, beams	Crankshafts, gears, fittings	Engine blocks, housings

Where extruded metal shows up

• Architectural framing. Window and door frames, curtain-wall mullions, solar-panel rails — almost universally 6063 aluminum through porthole dies, anodized or powder-coated after cutting.
• Heat sinks. The finned aluminum blocks on CPUs, LED fixtures, and power electronics are extruded with the fins as part of the die cross-section, then sawn to length. Fin aspect ratios up to roughly 20:1 are achievable.
• Transport structures. Railway carriage bodies, truck trailers, and shipping containers use large hollow aluminum extrusions; high-speed train bodies are friction-stir-welded from wide extruded panels. Aircraft stringers and seat tracks are extruded 7075 and 2024.
• Electrical. Copper and aluminum bus bars, conduit, and grounding sections are extruded for their long constant cross-section and good conductivity.
• Consumer and industrial. Aluminum T-slot framing (the 80/20 / Bosch Rexroth system), ladder rails, tent poles, telescopic tube, and the trim on nearly every appliance are extruded sections.

Common pitfalls when designing for extrusion

• Asymmetric wall thickness. Thin and thick regions in the same profile flow at different speeds, warping the extrudate and stressing the die. Keep walls as uniform as possible; add bearing-length correction in the die to balance flow.
• Sharp internal corners. Sharp corners concentrate stress on the die and choke metal flow. Generous radii extrude faster and last longer.
• Tongue ratios too high. A deep, narrow gap (a "tongue", e.g. a slot between two fins) leaves a slender die feature that can deflect or break. Keep the depth-to-width tongue ratio within die-maker limits (typically under ~3:1 for aluminum).
• Ignoring the circumscribing circle. A profile's size is limited by the press's container bore; a section that does not fit inside the maximum circumscribing circle simply cannot be run on that press.
• Forgetting the post-process. Extrusions stretch (to straighten and relieve stress), age-harden, cut, and finish after the press. Design tolerances must account for the as-extruded variation plus the stretch and aging, not the press exit alone.