Context and Motivation

Conventional ultrafine grained ferrite/cementite steels suffer from a fundamental shortcoming: despite their high strength, they exhibit very low strain hardening rates and consequently poor uniform elongation. The strategic replacement of cementite with martensite via intercritical annealing offers a elegant remedy—martensite restores strain hardening capacity while retaining the benefits of grain refinement. This paper systematically dissects how the processing parameters during intercritical annealing govern microstructure evolution in a plain C-Mn steel (0.17C, 1.63Mn, balance Fe) that had been pre-processed by large strain warm deformation to an ultrafine grained ferrite/spheroidized cementite state with a ferrite grain size of 0.84 μm.

Key Findings on Transformation Kinetics

The intercritical annealing parameters—holding temperature (710–750°C), holding time (2 s to 30 min), heating rate (0.25–100 K/s), and cooling rate (20–140 K/s)—were varied independently using dilatometry. The martensite fraction increases with temperature from 7.9 vol% at 710°C to 25.5 vol% at 750°C after 1 min holds, yet remains far below the equilibrium fraction of 45.9 vol% calculated via Thermo-Calc. This indicates that one minute is grossly insufficient to reach phase equilibrium. Holding time exerts a far stronger effect: 2 seconds already yield 6.9 vol% martensite, 1 min gives 24.3 vol%, and 30 min approach equilibrium at 37.6 vol%. The ferrite grain size, remarkably, stays nearly constant between 1.13 and 1.25 μm across all variations—a striking demonstration of grain size stability that sets this material apart from conventional DP steels.

The Crucial Role of Manganese Inheritance

The most interesting mechanistic insight concerns manganese. During the prior large strain warm deformation at 550°C, cementite becomes significantly enriched in Mn. This enrichment is inherited by the austenite that forms during intercritical annealing and is clearly detectable inside the martensite after quenching. At lower intercritical temperatures (near 710°C), phase transformation is controlled by sluggish Mn diffusion—the diffusion coefficient of Mn in austenite is seven orders of magnitude lower than that of carbon. At higher temperatures (above 730°C), carbon diffusion takes over as the rate-controlling mechanism. This explains why the martensite fraction at 710°C after 1 min is only 7.9 vol% compared to the equilibrium value of 29.6 vol%, while at 730°C the same holding time gives 24.3 vol% with equilibrium at 36.6 vol%. The inherited Mn enrichment also lowers carbon activity in austenite, further retarding growth kinetics.

Heating Rate Insensitivity and Nucleation Site Saturation

Unlike conventional cold-rolled DP steels where heating rate dramatically affects the final microstructure, this UFG starting material shows almost no sensitivity to heating rates between 0.25 and 100 K/s. The reason is nucleation site saturation: the warm-deformed microstructure provides such an enormous density of ferrite/cementite interfaces (cementite particles are homogeneously distributed and mostly located at grain boundaries) that even the slowest heating rate saturates available nucleation sites. Higher superheating simply cannot produce more austenite nuclei—there is no remaining untapped nucleation potential. This makes the processing route exceptionally robust.

Hardenability Paradox and Its Resolution

A surprising observation is the high hardenability despite the ultrafine grain size and absence of Cr or Mo. Typically, finer grains accelerate austenite decomposition because grain boundaries are the preferred nucleation sites for ferrite, pearlite, or bainite. However, cooling rates as low as 20 K/s still produce predominantly martensite rather than bainitic or pearlitic decomposition products. Two counteracting effects explain this: first, the Mn-enriched austenite inherits enhanced hardenability directly; second, during slower cooling, ferrite grows at the expense of austenite, enriching the remaining austenite in carbon. The ultrafine grain size makes this carbon enrichment particularly effective because diffusion distances are tiny. The two effects together compensate for the high grain boundary density that would otherwise accelerate austenite decomposition.

EBSD Methodology Innovation

A practical advance worth noting is the EBSD post-processing routine for distinguishing martensite from ferrite—both body-centered cubic and thus indistinguishable by conventional indexing. The method uses the bimodal distribution of Image Quality values (low IQ for martensite due to higher dislocation density, high IQ for ferrite) with an automatically defined threshold at the minimum between the two peaks. A subsequent neighbor criterion (a pixel belongs to martensite only if at least three neighboring pixels also have sub-threshold IQ) removes misindexed grain boundary pixels that would otherwise artificially inflate the martensite fraction. This is more objective than simple thresholding and avoids the problem of discarding too many grain boundary regions.

Grain Size Stability Mechanisms

The ferrite grain size increases only modestly from 0.84 μm in the initial ferrite/cementite condition to about 1.2 μm after intercritical annealing—regardless of holding time or temperature. Six mechanisms contribute to this exceptional stability: (1) pronounced recovery during warm deformation reduces the dislocation density, delaying recrystallization; (2) cementite particles effectively pin grain boundaries, with Mn refinement of cementite enhancing this effect; (3) growing austenite itself impedes ferrite grain boundary migration; (4) Mn in solid solution exerts solute drag; (5) Mn broadens the ferrite+austenite+cementite three-phase field where grain growth is strongly inhibited; and (6) Mn lowers the Ac₁ temperature, reducing grain growth kinetics at the annealing temperature.

Practical Processing Recommendation

The optimum intercritical annealing parameters identified are 730°C for 1 minute—sufficient to dissolve all cementite completely while keeping the martensite fraction at 24.3 vol%, well below the 30 vol% threshold where ductility typically deteriorates. The fastest cooling rate of 140 K/s preserves this martensite fraction. Slower cooling produces a finer, less percolated martensite distribution but at the cost of a slightly coarser ferrite grain size and reduced martensite fraction. The heating rate can be chosen freely within the investigated range without compromising the final microstructure.