Work Hardening: Why Stainless Punishes a Light Cut

Process / 16 Sept 2026

There is a failure mode specific to austenitic stainless that catches shops which do not run it often. The part goes well until a finishing pass, then the surface starts tearing, the tool wears out in minutes, and every attempt to fix it by going lighter makes it worse.

That is work hardening, and understanding it explains most of what is odd about machining SUS304 and SUS316.

What is physically happening

Austenitic stainless is face centred cubic, non magnetic, and at room temperature it is metastable. The austenite is not the structure the alloy would prefer at that temperature, it is the structure that was frozen in by solution annealing.

Deform it and some of it transforms to martensite. Strain induced martensite is hard and brittle. It is the same transformation that hardens a carbon steel by quenching, except here the trigger is mechanical rather than thermal.

Some numbers to give it scale. Annealed SUS304 sits around 200 HV. A heavily work hardened surface layer can reach 400 HV or more, and that layer can be a few hundredths of a millimetre deep. The next pass has to cut through it.

Two other properties make it worse:

Low thermal conductivity. SUS304 conducts at roughly 16 W per metre kelvin, about a third of plain carbon steel. Heat that should leave in the chip stays at the cutting edge instead.

High ductility and a tendency to gum. The chip wants to stay attached. Built up edge forms on the rake face, tears away, and takes coating with it.

The trap: going lighter makes it worse

This is the counterintuitive part, and it is where shops new to stainless lose parts.

A tool that is cutting below the depth of the previously hardened layer is rubbing, not cutting. Rubbing deforms the surface without removing it, which creates more hardened material, which makes the next pass worse still. The instinctive response to a poor surface is to reduce the feed and take a lighter cut, which accelerates exactly the problem it was meant to fix.

The rule shops that run stainless live by: always cut under the previous pass’s hardened layer. Depth of cut is a minimum, not a maximum. A finishing pass at 0.05 mm on SUS316L is often worse than one at 0.3 mm.

Related consequences:

Never dwell. A tool that stops while still in contact work hardens the spot it stopped on. On a lathe, a pause at the end of a cut leaves a hard ring
Never re-cut the same surface at the same depth. A spring pass on stainless is a hardening pass
Change tools on a schedule. A worn edge rubs. Waiting for visible failure means the last twenty parts were made by a rubbing tool
Enter and leave cleanly. Rolling into a cut is better than plunging and reversing

What a shop does about it

The practical measures, briefly, so you can recognise a shop that knows them:

Sharp, positive geometry. Honed edges, positive rake, polished flutes. A tool designed for steel will rub on stainless.

Heavier feed than instinct. Feed per tooth high enough to get under the hardened layer, always.

Lower surface speed than carbon steel, but not a crawl. Typically 30 to 50 % below what the same tool would run in mild steel.

Coolant that reaches the cut. High pressure through spindle where the machine supports it. Flood coolant that lands beside the cut does nothing for heat at the edge.

Rigidity. Deflection means variable depth of cut, and the shallow part of a variable cut is a rubbing cut.

Finishing stock planned in. Typically 0.3 to 0.5 mm left on diameters, so the finish pass cuts fresh metal rather than skin.

What this means for your drawing

This is where a buyer has more influence than they expect. Several drawing features force a shop into light cuts, and each one is a work hardening risk.

Thin walls. A wall that deflects under a normal cut forces a light pass, which hardens, which makes the next pass worse. Below roughly 1 mm on stainless you are into specialist territory. If a wall is thin for a real reason, keep it. If it is thin because the model was scaled from an aluminium part, it is costing you.

Deep narrow pockets and slots. A long tool deflects. Deflection means the cut varies, and the shallow part of that variation rubs. Opening a slot by a millimetre can let the shop use a shorter, stiffer tool.

Interrupted cuts. Every re-entry is an impact on a hardened surface. A cross hole through a diameter that is then turned is a known difficulty.

Sharp internal corners. They force a small tool, which means a light cut and a slow feed. A 3 mm radius instead of a 1 mm radius can halve the time in a pocket.

Tight tolerances applied everywhere. A blanket fine general tolerance class forces finishing passes on faces that did not need them. Tolerance the three features that matter and let the rest run at medium class. See the general tolerance guide.

Tight surface finish applied everywhere. Same logic. Ra 1.6 is free, Ra 0.4 is an operation. See JIS surface finish symbols.

Grade choice is the biggest lever

If the service allows it, changing the grade removes the problem rather than managing it.

Grade	Work hardening	Machinability index
SUS303	Low, sulphides break the chip	78
SUS304	High	45
SUS316 / 316L	High	36
SUS430	Low, ferritic, no transformation	54
SUS410, SUS420 annealed	Low, martensitic	45 to 54

For a turned part that is never welded and never wet, SUS303 halves the problem and the cycle time at roughly the same bar price. For a dry indoor part, SUS430 does something similar. Both are decisions that belong to whoever owns the design, and both are worth raising before the part goes into production. The comparisons are in SUS303 vs SUS304 vs SUS316 and SUS430 vs SUS304.

One more consequence: magnetism

The martensite created by work hardening is magnetic. A SUS304 part that was non magnetic as bar can be measurably magnetic after heavy machining, particularly at the surface and around cold worked features.

Usually this is cosmetic. Occasionally it is not. If the part sits near a sensor, inside a vacuum chamber or in an instrument, specify the limit on the drawing, for example relative permeability below 1.05, and expect the shop to plan a solution anneal or a lighter finishing strategy. Discovering it after delivery is expensive, and the drawing is the only place the requirement is binding.

Have a stainless part where the geometry is fighting you? Get a quote. The quote tells you which features are driving the price before you commit to them. Within 24 business hours.

Questions

Why does stainless work harden more than carbon steel?

Austenitic stainless has a metastable austenite structure. Plastic deformation converts some of it to martensite, which is much harder. Carbon steel work hardens too, but it has no phase transformation waiting to happen, so the effect is far smaller.

Does a slower feed help?

No, it usually makes it worse. A light feed lets the tool rub rather than cut, which deforms the surface without removing it, which is exactly what creates the hardened layer.

Which grades are worst?

The austenitics: SUS304, SUS316 and SUS316L. SUS303 is far better because the sulphide inclusions break the chip. The ferritic SUS430 and the martensitic grades barely work harden at all.

Is this something a buyer can influence?

Partly. Thin walls, deep narrow pockets, interrupted cuts and blanket tight tolerances all force the light passes that cause the problem. Removing an unnecessary one from the drawing is worth more than any cutting parameter.

Have a drawing for a stainless part?

Send it and an engineer reads the title block, quotes it within 24 business hours, and if you order we place, manage, inspect and ship the parts on one PO and one invoice.

Get a quote