Fushun Metal · Technical Notes
By Harris, technical staff at Fushun Metal
Forget the Annealing Temperature: Why the Slow-Cool Stop Decides DP600
Most DP600 process debates are stuck on the wrong knob. Ask any mill engineer how to tune a DP steel and the first number out of their mouth is the intercritical annealing temperature. That is understandable — annealing sets how much austenite forms — but it is only half the story, and it is the half that gets all the press. The decisive, under-rated variable is what happens in the next few degrees of the descent: the slow-cooling stop temperature, the plateau between the two-phase anneal and the final fast quench.
Let us fix the grade vocabulary first, because DP600 crosses borders under different labels. This study uses a 600 MPa-class cold-rolled dual-phase steel, the C-Si-Mn workhorse for car bodies. The same family is sold in North America as AISI/USAMP DP590 / DP600, and in Europe under EN 10268 as DP600 (material grade 1.0338). Whatever the number on the certificate, the metallurgy is identical: a soft ferrite matrix studded with hard martensite islands, which is what buys the low yield point, high tensile strength, high initial work-hardening and forming capability that make it a core AHSS for lightweighting.
The chemistry is built for carbon partitioning
The tested steel is deliberately simple, and every element earns its place during the slow-cool step rather than the anneal.
| C | Si | Mn | P | S | Cr | Nb | Fe |
|---|---|---|---|---|---|---|---|
| 0.10 | 0.35 | 1.85 | ≤0.020 | ≤0.015 | 0.15 | 0.03 | balance |
Silicon suppresses cementite so carbon has nowhere to go but into the austenite; manganese both raises hardenability and slows diffusion-controlled products; niobium pins the grain. None of that pays off at the anneal — it pays off during the controlled hold where carbon is allowed to migrate. On the Gleeble the alloy measured Ac₁ = 738 °C and Ac₃ = 849 °C.
What the anneal really does — and where its job ends
Heating into the two-phase field dissolves part of the ferrite, and the higher you go, the more austenite you create. That relationship is monotonic and well-known, and it is exactly why people stop thinking here.
| Annealing temp (°C) | Ferrite | Austenite |
|---|---|---|
| 760 | 64 | 36 |
| 780 | 55 | 45 |
| 800 | 43 | 57 |
| 820 | 37 | 63 |
820 °C was chosen as the reference anneal because it gives the fullest austenite (63 %). But note what the table cannot tell you: how much carbon is inside that austenite, and how stable it is. Those answers are set entirely downstream, in the slow-cool window.
The slow-cool window is where carbon is redistributed
Cooling from the anneal toward the quench is not a passive transition. Between roughly 700 and 620 °C the steel sits in the proeutectoid ferrite formation zone: new ferrite nucleates, and because ferrite holds almost no carbon, the rejected carbon is pushed into the remaining austenite. This is carbon partitioning, and it is the whole point of the hold. Enriched austenite is stable austenite — stable enough to survive to the quench and transform to martensite, yet not so unstable that it collapses into pearlite or bainite.
Stop too high and the partitioning barely happens: less proeutectoid ferrite, less carbon enrichment, weaker austenite. Stop too low and you over-produce ferrite and lock in a coarse, blocky martensite. The optimum is a narrow ledge, and it is invisible to anyone who only reads their annealing setpoint.
The bainite competition everyone forgets
Here is the sharp claim. Raising the slow-cool stop temperature does not simply lower ferrite — it changes what the hard phase is. At a 620 °C stop, the micrograph is clean: polygonal ferrite plus blocky martensite at the grain boundaries. At a 700 °C stop, blocky martensite thins out and lath-type bainite appears, with dark acicular regions in the scanning-electron images. Under-enriched austenite cannot hold out for a pure martensitic transformation, so it decomposes to bainite during the fast cool. Martensite and bainite are in direct competition for the same austenite, and the slow-cool stop temperature picks the winner.
The numbers: 40 °C moves every phase
Quantified, the effect is unmistakable. The initial (undissolved) ferrite stays pinned at 37 % — that is fixed by the anneal. Everything that changes is born in the slow-cool window.
| Slow-cool stop (°C) | Initial ferrite | Proeutectoid ferrite | Total ferrite | Martensite / bainite |
|---|---|---|---|---|
| 620 | 37 | 45 | 82 | 18 |
| 640 | 37 | 42 | 79 | 21 |
| 660 | 37 | 38 | 75 | 25 |
| 680 | 37 | 30 | 67 | 31 |
| 700 | 37 | 26 | 63 | 37 |
Across a 40 °C band, proeutectoid ferrite falls from 45 % to 26 % while the hard-phase fraction climbs from 18 % to 37 %. That is a full phase re-architecture achieved without touching the annealing temperature at all.
Mechanical properties follow a non-obvious optimum
Strength does not rise monotonically with the hard phase, and ductility does not fall monotonically either. Both curves turn over, and they turn over at the same place.
| Slow-cool stop (°C) | Yield strength (MPa) | Tensile strength (MPa) | Total elongation (%) |
|---|---|---|---|
| 620 | ≈400 | ≈715 | ≈22 |
| 640 | ≈397 | ≈713 | ≈24 |
| 660 | 387 | 712 | 27 |
| 680 | ≈400 | ≈725 | ≈22 |
| 700 | ≈415 | ≈735 | ≈21 |
The sweet spot is the 660 °C stop: yield 387 MPa, tensile 712 MPa, and a peak total elongation of 27 %. Below it, abundant blocky martensite hardens the steel but hurts ductility; above it, strength creeps back up as the hard-phase fraction grows, but the loss of ferrite and the onset of bainite pull elongation back down. The reason 660 °C wins is the same competition described above — a little bainite narrows the ferrite/martensite hardness gap, resists interface decohesion, and lets deformation run longer before fracture.
The takeaway
For anyone specifying or producing DP600, DP590 or EN 1.0338, the lesson is to stop treating the cooling ramp as a formality. The anneal decides how much austenite exists; the slow-cool stop decides its carbon content, its stability, whether the hard phase ends up martensite or bainite, and ultimately the strength-ductility balance you sign off on. A 40 °C swing in a single hold moved the hard-phase fraction by 19 points and elongation by 6 points.
So the next time a DP600 coil forms short or a spot-weld heat-affected zone cracks, do not reach first for the furnace setpoint. Look at the soak between the anneal and the quench. That quiet, unglamorous plateau is where the steel is actually made.
