Hot rolling is far more than a shaping operation — it is the primary microstructure engineering process for structural steels, linepipe grades, heavy plates, and structural sections. The rolling schedule: the sequence of reduction passes, inter-pass times, rolling temperatures, strain rates, and post-rolling cooling rates, determines the final ferrite grain size, strength, toughness, weldability, and anisotropy of the finished product. Modern thermomechanical controlled processing (TMCP) exploits the precise interaction of deformation, recovery, recrystallisation, precipitation, and phase transformation to achieve combinations of strength and toughness that are simply impossible by any other processing route — and at lower alloy content, making the steel more weldable to boot.

Metallurgical Events in Hot Rolling: DRX, SRX, and Grain Growth

At hot rolling temperatures (900–1250°C), steel is in the fully austenitic condition. During rolling, three distinct metallurgical processes operate in sequence and competition, each governing the austenite microstructure handed to the subsequent stage. Understanding the kinetics of each process is essential for mill schedule design.

1. Dynamic Recrystallisation (DRX)

Dynamic recrystallisation occurs during a rolling pass when the accumulated strain exceeds a critical value ε_c. New, dislocation-free austenite grains nucleate — typically at prior grain boundaries, deformation bands, and twin intersections — and grow while deformation continues simultaneously. DRX is characterised by the Zener-Hollomon parameter Z:

Zener-Hollomon parameter:
  Z = ε̇ × exp(Q_def / RT)

Where:
  ε̇    = strain rate (s⁻¹)
  Q_def = activation energy for hot deformation of austenite ≈ 300 kJ/mol
  R     = 8.314 J/(mol·K)
  T     = absolute temperature (K)

Critical strain for DRX onset:
  ε_c ≈ 0.8 × ε_p    (typically ε_c ≈ 0.4–0.8, depending on Z)
  where ε_p = peak strain at flow stress maximum

DRX-recrystallised grain size (Sellars equation):
  d_DRX = A × Z^(-n)   (n ≈ 0.15–0.25 for austenite)
  → Higher Z (faster/colder) → finer DRX grain size
  → Example: Z ≈ 10¹⁵: d_DRX ≈ 40 µm; Z ≈ 10¹⁸: d_DRX ≈ 15 µm

DRX is the dominant softening mechanism in high-temperature passes (>1100°C) on continuous tandem mills where the inter-pass interval is too short for SRX to complete. It produces a fine, near-equiaxed recrystallised grain structure with relatively low dislocation density. In production plate mills with longer inter-pass times, SRX typically dominates over DRX even in the roughing stages.

2. Static Recrystallisation (SRX)

Static recrystallisation occurs during the inter-pass interval, driven by the stored deformation energy (dislocations) introduced by the pass. New grains nucleate at deformed grain boundaries and grow at the expense of deformed grains. SRX kinetics follow the Johnson-Mehl-Avrami equation:

Static recrystallisation kinetics (JMAK equation):
  X_SRX = 1 − exp[−0.693 × (t/t_50)^n]   (n ≈ 1.5–2.0 for austenite)

Time for 50% SRX (Sellars / Hodgson-Gibbs type):
  t_50 = A × d₀^p × ε^(-q) × ε̇^(-r) × exp(Q_SRX / RT)

Typical parameters for plain C-Mn austenite:
  Q_SRX ≈ 300 kJ/mol; p ≈ 2; q ≈ 4; r ≈ 0
  → At 1100°C, ε=0.3, d₀=100µm: t_50 ≈ 2–5 s  (SRX fast — complete in one pass gap)
  → At 950°C, ε=0.3, d₀=100µm:  t_50 ≈ 50–200 s  (SRX very slow — incomplete)
  → With 0.035% Nb at 950°C: t_50 ≈ 1000–10000 s  (SRX essentially suppressed)

SRX grain size:
  d_SRX = C × d₀^a × ε^(-b)   (independent of temperature in many empirical forms)
  → Smaller prior grain size and higher strain → finer SRX grain size

Progressive SRX through the roughing pass sequence progressively refines the austenite grain size from the as-reheated size (~150–250 µm) down to ~50–80 µm at the end of roughing — the starting structure for the critical finish rolling stage. The completeness of SRX between roughing passes is why roughing must be conducted above T_nr.

3. Metadynamic Recrystallisation (MDRX)

A third mechanism — metadynamic recrystallisation — occurs in the very short interval immediately after a DRX pass. Nuclei formed during DRX continue to grow rapidly during the inter-pass gap, driven by stored energy, without the need for a conventional nucleation event. MDRX kinetics are much faster than SRX (no incubation period) and produce a finer grain size than SRX from the same prior structure. In tandem hot strip mills with inter-pass times of 0.1–1 second, MDRX is the dominant inter-pass softening mechanism for roughing passes where DRX has been triggered.

4. Grain Growth

After recrystallisation is complete, the recrystallised grains grow to minimise grain boundary energy at a rate governed by:

Normal grain growth (parabolic law):
  d² − d₀² = K_gg × t
  K_gg = K₀ × exp(−Q_gg / RT)   (Q_gg ≈ 400 kJ/mol for austenite)

Zener-pinning limit for grain coarsening (precipitates present):
  d_limit = (4r) / (3f_v)

Where:
  r   = precipitate radius (m)
  f_v = precipitate volume fraction

Effect of microalloying on grain coarsening temperature:
  TiN (dissolves at ~1450°C):  pins grains up to slab reheating temperature
  NbC/N (dissolves at 1050–1200°C): pins grains during high-T roughing
  AlN (dissolves at ~1100°C):   grain boundary pinning in Al-killed steels
  VC (dissolves at ~950°C):     too fine / low-T to pin during hot rolling

Typical austenite grain sizes:
  As-reheated (1200°C, no Ti): 150–250 µm
  As-reheated (1200°C, 0.015% Ti + 0.035% Nb): 80–120 µm
  After roughing (SRX complete): 40–80 µm
  After finish rolling below Tnr (pancaked): aspect ratio 5–10:1, 
         effective nucleation diameter 5–20 µm

The Non-Recrystallisation Temperature (T_nr)

T_nr is the most critical metallurgical parameter in TMCP schedule design. Below this temperature, niobium in solid solution (and as fine NbC/N precipitates on austenite grain boundaries and subgrain boundaries) so strongly retards the nucleation and early growth of SRX that recrystallisation is effectively suppressed during the inter-pass interval. The primary T_nr model is the Boratto et al. (1988) equation:

Boratto et al. (1988) T_nr equation:
  T_nr (°C) = 887 + 464×C + (6645×Nb − 644×√Nb)
                          + 732×V − 230×√V
                          + 890×Ti + 363×Al − 357×Si

Typical T_nr values:
  Plain C-Mn steel (0.15C, 1.5Mn, no Nb):      T_nr ≈ 800–850°C
  Low Nb (0.015Nb):                              T_nr ≈ 900°C
  Standard HSLA (0.07C, 0.035Nb, 0.015Ti):      T_nr ≈ 940–960°C
  High Nb (0.06Nb):                              T_nr ≈ 970–980°C
  Nb+V (0.04Nb, 0.06V):                         T_nr ≈ 980°C

Mechanism of Nb retardation of SRX:
  1. Solute drag: Nb atoms segregate to moving austenite grain boundaries
     and sub-boundaries, exerting a drag force opposing boundary migration
     → Proportional to Nb solute content × boundary velocity
  2. Precipitate pinning (Zener): Fine NbC/N particles (2–20 nm) on boundaries
     resist boundary movement with force proportional to f_v/r
  3. Combined effect is dramatically longer t_50 for SRX below T_nr,
     effectively making SRX kinetically inaccessible during normal pass gaps

Austenite Pancaking: The Core of TMCP

When rolling below T_nr with suppressed SRX, each pass flattens and elongates the austenite grains in the rolling direction (RD) and reduces their thickness in the normal direction (ND). The result is a “pancaked” austenite microstructure with aspect ratios of 5–10:1 (occasionally higher with heavy reductions) and dramatically increased microstructural surface area per unit volume S_v.

Pancaked austenite microstructural parameters:

Grain boundary area per unit volume after pancaking:
  S_v (m⁻¹) ≈ 2/d_ND + 1/d_RD   (for idealized flattened ellipsoidal grains)
  where d_ND = grain thickness in normal direction
        d_RD = grain dimension in rolling direction

Effect on ferrite nucleation (Militzer model):
  N_α (nuclei/m³) ∝ S_v × (1 + f_bands)
  where f_bands = fraction of grain interior with deformation bands
        (adds 20–40% additional nucleation sites in heavily deformed austenite)

Example comparison (same steel composition):
  Recrystallised austenite: d ≈ 60 µm, S_v ≈ 33,000 m⁻¹
    → Ferrite grain size after air cool: d_α ≈ 15–20 µm (ASTM 8–9)
  
  Pancaked austenite: d_ND ≈ 10 µm, d_RD ≈ 80 µm, S_v ≈ 215,000 m⁻¹
    → Ferrite grain size after ACC: d_α ≈ 5–8 µm (ASTM 11–13)

Hall-Petch yield strength increase from grain refinement:
  σ_y = σ_0 + k_y × d^(-0.5)   (k_y ≈ 0.6 MPa·m^0.5 for ferrite)
  Δσ (20→6 µm) = 0.6 × (1/√(6×10⁻⁶) − 1/√(20×10⁻⁶))
              = 0.6 × (408 − 224) = 110 MPa

The physical mechanism of ferrite grain refinement from pancaked austenite is a straightforward application of nucleation theory: ferrite nucleates at austenite grain boundaries with a nucleation rate proportional to the total available boundary area S_v. More nuclei growing simultaneously means each nucleus has less volume to fill before impingement with neighbours — producing a finer final grain size. Deformation bands within the pancaked grains (regions of locally high dislocation density that form during heavy reductions below T_nr) provide additional intragranular nucleation sites, further reducing the effective grain size.

Accelerated Cooling (ACC): Extending the TMCP Benefit

Accelerated cooling begins immediately after the final finish rolling pass, typically from temperatures of 800–850°C (above Ar3 to maintain full austenite before cooling). Water spray banks apply cooling rates of 10–50°C/s to the plate surface. The cooling rate and stop temperature are the primary parameters controlling the final transformation microstructure.

ACC design parameters and microstructure outcomes:

Cooling rate (°C/s) → Microstructure → Typical YS (X65 composition):
  Air cool (~0.5°C/s):    Polygonal ferrite + pearlite        ~420 MPa
  Mild ACC (5–10°C/s):    Fine polygonal + acicular ferrite   ~460 MPa
  Standard ACC (15–25°C/s): Acicular ferrite + bainite        ~500 MPa
  Heavy ACC (25–50°C/s):  Bainite dominated                  ~560 MPa
  Direct quench (DQ):     Martensite + bainite               ~690+ MPa (requires temper)

ACC stop temperature → Microstructure:
  650–700°C: Fine polygonal ferrite + pearlite (X52–X60 range)
  550–650°C: Acicular ferrite + granular bainite (X65 target)
  450–550°C: Lower bainite + martensite-austenite constituents (X70–X80)
  <400°C:   Martensite-bainite (DQ+T required for X100+)

Surface-centre temperature gradient:
  For 20mm plate, ACC at 20°C/s:
    Surface: ~25°C/s → bainite
    Quarter-thickness: ~15°C/s → acicular ferrite + bainite
    Centre: ~8°C/s → fine polygonal ferrite + acicular ferrite
  → Through-thickness property variation must be within specification limits

Microalloying Elements: Nb, V, Ti, and B in TMCP Steels

The effectiveness of TMCP depends critically on the presence of one or more microalloying elements at concentrations of 0.01–0.12 wt%. These elements act at different temperature stages through different mechanisms:

Solubility Products and Precipitation Temperature

The temperature at which each carbide or nitride precipitates in austenite determines its role in the rolling process. The solubility product K_sp(T) for NbC in austenite (one of the most important precipitation reactions in TMCP) is:

Solubility products in austenite:

NbC:  log([Nb][C]) = 2.26 − 6770/T      (T in Kelvin, concentrations in wt%)
NbN:  log([Nb][N]) = 3.70 − 10800/T
TiN:  log([Ti][N]) = 0.32 − 8000/T     → Dissolves above ~1450°C
TiC:  log([Ti][C]) = 5.33 − 10475/T

Example: X70 steel (0.07C, 0.035Nb, 0.010N, 0.012Ti)
  All Nb in solution at 1200°C (reheating target)
  NbC starts precipitating at T ≈ 1020°C (coarsening of fine NbC begins)
  TiN remains as undissolved particles throughout entire hot rolling cycle
  → Ti pins boundaries at ALL temperatures; Nb in solution retards SRX below T_nr

Reheating temperature selection:
  Too low (<1050°C): Nb remains as undissolved coarse NbC → insufficient dissolved
                     Nb for T_nr effect → poor pancaking → low toughness
  Too high (>1220°C): TiN starts to dissolve → austenite grain coarsening →
                      coarse starting grain → final grain size too large
  Optimum:            1150–1180°C for most Nb-HSLA linepipe steels

Strengthening Mechanisms in TMCP Steel

The yield strength of a TMCP structural steel is the sum of multiple strengthening contributions. For an X65 linepipe steel (minimum yield strength 450 MPa), a typical breakdown is:

Strengthening mechanisms — additive rule (Gladman):
  σ_y = σ_0 + Δσ_SS + Δσ_HP + Δσ_ppt + Δσ_disl

Grain refinement (Hall-Petch):
  Δσ_HP = k_y × d^(-0.5)   (k_y ≈ 0.6 MPa·m^0.5 for ferrite)

Solid solution strengthening (linear additivity):
  Δσ_SS ≈ 32×%Si + 37×%Mn + 11×%Mo + 18×%Cu + 2.5×%Ni  (MPa, approximate)

Precipitation hardening (Orowan-Ashby):
  Δσ_ppt = M × 0.4Gb/π × 1/(L√(f_v)) × ln(r/b)
  Simplified for NbC:  Δσ_ppt ≈ 700×%Nb^0.5  (MPa, approximate rule of thumb)

TMCP Grade Properties and Linepipe Steel Evolution

Acicular Ferrite: The Ideal TMCP Microstructure

Acicular ferrite (AF) is the microstructure that delivers the optimal combination of strength and toughness in TMCP linepipe and structural plate steels. It forms by an intragranular nucleation mechanism on non-metallic inclusions (primarily titanium oxide or calcium-alumino-silicate inclusions engineered by tundish treatment) and at austenite grain boundaries and deformation bands simultaneously. The resulting microstructure consists of randomly oriented, interlocking fine ferrite laths (1–5 µm wide, 5–20 µm long) with no preferred orientation and high angle grain boundaries between adjacent laths.

The key toughness advantage of acicular ferrite over polygonal ferrite or bainite is the extremely tortuous crack path required for brittle fracture: with high-angle boundaries at every 2–5 µm, cleavage cracks must repeatedly renucleate across boundary misorientations, dramatically increasing fracture energy and lowering the ductile-brittle transition temperature (DBTT) to −60°C or below. For a detailed treatment of cleavage fracture and the role of crystallography in fracture, see the Charpy Impact Test and Grain Boundaries articles.

Texture and Anisotropy in Hot-Rolled Plate

Hot rolling in the non-recrystallising region not only refines grain size but also develops crystallographic texture — a preferred orientation distribution of ferrite grains that produces measurable mechanical anisotropy in the finished plate. The primary hot-rolling deformation texture components in steel plate are:

• γ-fibre ({111}⟨uvw⟩): Develops preferentially during finish rolling below T_nr. Grains with {111} planes parallel to the plate surface are the ideal texture for deep drawing (high Lankford r-value); they also contribute to improved through-thickness toughness by minimising the number of grain boundaries with low-misorientation-angle facets favouring through-thickness cleavage crack propagation.
• α-fibre ({hkl}⟨110⟩): Develops during both DRX and SRX above T_nr. The ⟨110⟩ component aligns easy magnetisation directions with the rolling direction — relevant for electrical steels.
• Rotated cube ({001}⟨110⟩): Produces the lowest yield strength in the transverse direction (TD) and highest anisotropy; undesirable in structural plate requiring isotropic properties.

The through-thickness anisotropy (Z-direction properties) is particularly critical for offshore structural steels loaded in the thickness direction (e.g., crane pedestals, ring-welded connections). This is mitigated by: (1) ultra-low sulphur content (<0.003 wt%) to minimise MnS inclusion stringers; (2) calcium treatment to spheroidise residual sulphide inclusions; and (3) ensuring sufficient cross-rolling (rolling in both RD and TD) to avoid excessive pancaking in one direction only. Z-quality steel (minimum 35% reduction of area in the thickness direction, e.g. S355K2Z35 per EN 10164) specifies through-thickness ductility requirements to prevent lamellar tearing in welded structural details.

Texture, Anisotropy, and Z-Quality Considerations

Rolling in the non-recrystallising region develops a pronounced crystallographic texture in both the austenite and the resulting ferrite. The dominant texture component in finish-rolled and ACC-cooled linepipe steel is the γ-fibre ({111}⟨uvw⟩), which gives a high in-plane Lankford r-value and relatively isotropic in-plane properties — beneficial for pipe forming by UOE or spiral welding. However, the same rolling process that creates the pancaked structure introduces mechanical fibering (elongated MnS and oxide inclusions aligned in the RD) and texture anisotropy that reduces through-thickness (Z-direction) ductility.

For offshore jacket structures and other through-thickness-loaded connections, Z-quality steel (EN 10164, typically Z35 grade: minimum 35% reduction of area in thickness tensile test) is specified. This requires:

• Sulphur below 0.005 wt% (ideally below 0.002 wt%) to minimise MnS stringers.
• Calcium treatment (Ca/S > 1.5 by weight) to transform MnS to calcium-alumino-silicate inclusions that remain nearly spherical through rolling rather than elongating as stringers.
• Cross-rolling or multiple-direction rolling to reduce texture severity and improve through-thickness isotropy.
• Clean steel practice (low total oxygen, low P) to minimise all non-metallic inclusions.

For the underlying thermodynamics of austenite transformation during accelerated cooling, see the Austenite in Steel article and the CCT Diagram guide, which explains how the pancaked austenite CCT diagram is constructed and used to select ACC parameters. The resulting bainite and acicular ferrite microstructures are treated in depth in Bainite Microstructure in Steel. For the weldability implications of TMCP CE_IIW and HAZ microstructure, see HAZ Microstructure and Hydrogen-Induced Cracking.