Widmanstätten Ferrite: Formation, Morphology, and Effect on Toughness

Widmanstätten ferrite (WF) is a coarse, plate-shaped or needle-shaped ferrite morphology that grows into austenite along specific crystallographic directions at intermediate cooling rates, producing a microstructure that is thermodynamically inferior to equiaxed ferrite-pearlite and dramatically inferior in toughness to acicular ferrite or bainite. It is routinely encountered in the coarse-grained heat-affected zone (CGHAZ) of high-heat-input steel welds, in cast steel sections, and in any steel that has been cooled slowly through the 600–750°C temperature range after exposure to coarse prior austenite grain sizes. Understanding WF formation, its structural origins, and its suppression is essential for weld procedure qualification, PWHT design, and materials selection in structural, pressure vessel, and pipeline applications.

Historical Context and Nomenclature

The term “Widmanstätten structure” derives from Alois von Widmanstätten, who in 1808 observed that polished and acid-etched cross-sections of iron meteorites revealed long, interlocking bands of two iron-nickel phases (kamacite and taenite) arranged in crystallographically related directions within the original solidification grains. The same geometrically regular, crystallographically controlled plate morphology was later identified in the ferrite formed during solid-state austenite decomposition in steels, and the name was adopted by analogy.

In modern metallurgical terminology, Widmanstätten ferrite refers specifically to the proeutectoid ferrite morphology that grows as plates or laths into austenite along well-defined crystallographic habit planes, distinct from: allotriomorphic ferrite (equiaxed grains nucleating on grain boundaries without crystallographic growth constraint); idiomorphic ferrite (equiaxed grains nucleating intragranularly); acicular ferrite (intragranular plates nucleating on inclusions with random mutual impingement); and bainite (forming below the bainite start temperature Bs with a specific carbon-supersaturation driving force).

Thermodynamic and Kinetic Framework

WF formation occupies a specific field on the continuous cooling transformation (CCT) diagram, bounded thermodynamically by the Tγs temperature above and the bainite start temperature Bs below, and kinetically by the cooling rate range that allows sufficient carbon diffusion for plate lengthening without permitting complete equiaxed ferrite impingement.

The Tγs (Widmanstätten Start) Temperature

Tγs is the highest temperature at which WF can nucleate and grow from austenite of a given composition. It is always below Ae3 — thermodynamic equilibrium requires a temperature below Ae3 for any ferrite nucleation — but is significantly above Bs. The gap between Ae3 and Tγs represents the temperature window where only equiaxed proeutectoid ferrite can form. Bhadeshia (1992) derived the following empirical expression:

Tγs (°C) = 1290 − 25.8Mn − 37.7Ni − 37.7Cr − 41.1Mo
              − 230√C    (compositions in wt%)

WF formation field:  Tγs > T > Bs

Bainite start (Bs):
  Bs (°C) = 830 − 270C − 90Mn − 37Ni − 70Cr − 83Mo

Critical carbon limit:
  At [C] ≈ 0.53 wt% (plain C steel): Tγs = Bs
  Above this carbon, WF field disappears entirely.
  All transformation below Ae3 goes directly to bainite or pearlite.

Example (0.15C-1.4Mn-0.3Ni-0.1Cr, wt%):
  Tγs ≈ 1290 − 25.8(1.4) − 37.7(0.3) − 37.7(0.1) − 230√0.15
       ≈ 1290 − 36.1 − 11.3 − 3.8 − 89.1
       ≈ 749°C
  Bs   ≈ 830 − 270(0.15) − 90(1.4) − 37(0.3) − 70(0.1)
       ≈ 830 − 40.5 − 126 − 11.1 − 7.0
       ≈ 645°C
  WF field: 749 → 645°C for this steel composition.

Transformation Temperature Field Diagram

Growth Mechanism: The Displacive-Diffusive Hybrid

WF growth is best described as a hybrid mechanism that is neither fully diffusive (as in allotriomorphic ferrite) nor fully displacive (as in martensite). The plate interface advances by a ledge mechanism: individual ledges, typically 5–50 nm high and separated by flat terrace segments, migrate laterally along the habit plane. The terrace segments are coherent with the austenite, allowing rapid interface advance by a shear-like motion. Carbon rejected from the growing ferrite (which is virtually carbon-free at transformation temperatures) must diffuse away from the advancing interface through the austenite ahead of the plate tip, creating a carbon-enriched diffusion field that limits the plate lengthening rate.

The plate tip advances faster than the broad faces (sides) because the carbon diffusion geometry at a sharp tip provides a steeper concentration gradient and faster removal of rejected carbon. This anisotropic growth produces the characteristic long, thin plate geometry with a high length-to-width aspect ratio, typically 5:1 to 20:1 in steels.

Plate lengthening rate (simplified Trivedi-Pound model):

v ≈ αC / ρ

where:
  v       = plate tip velocity (m/s)
  αC     = carbon diffusivity in austenite at transformation temperature
  ρ       = plate tip radius (~10–50 nm)

Carbon diffusivity in austenite at 700°C:
  DC(γ) ≈ 2 × 10⁻¹¹ exp(−142,000 / RT) m²/s
           ≈ 1.5 × 10⁻¹³ m²/s at 700°C

  At typical ρ = 20 nm:  v ≈ 1.5×10⁻¹³ / 20×10⁻³ ≈ 75 µm/s
  A 200 µm plate could form in ~2.7 seconds at this rate —
  much faster than equiaxed grain growth by diffusion alone.

Morphological Classification

WF is categorised by its nucleation site and spatial arrangement. The two primary subdivisions — primary and secondary — have distinct microstructural signatures and slightly different property consequences.

Primary Widmanstätten Ferrite (WFα)

Primary WF nucleates directly on the prior austenite grain boundary without any preceding allotriomorphic ferrite layer. This requires that the grain boundary be in a specific crystallographic orientation relative to the austenite lattice on one side — only boundaries where the K-S relationship can be established with one of the adjacent grains will nucleate primary WF. Primary WF forms at higher temperatures within the WF field (close to Tγs) and at lower carbon contents (typically below 0.10 wt% C), where the driving force for WF is high relative to the driving force for allotriomorphic ferrite.

In optical micrographs etched with 2% nital, primary WF appears as bright ferrite plates radiating from the grain boundary into the grain interior with no visible allotriomorph layer. The plates are typically 10–30 μm wide and may traverse the entire prior austenite grain, producing a “spike” or “fan” arrangement. Very low-carbon steels (<0.06 wt% C) can exhibit primary WF in castings and weld HAZ with prior austenite grain sizes exceeding 500 μm.

Secondary Widmanstätten Ferrite (WFαs)

Secondary WF is by far the more common morphology in engineering steels. It nucleates on an existing grain-boundary allotriomorph (GBA) ferrite layer that has already formed at the austenite grain boundary during initial cooling through the Ae3–Tγs temperature window. The secondary WF plates grow from the allotriomorph into the austenite interior as parallel or near-parallel side-plates — producing the characteristic “Widmanstätten structure” that defines the morphology in most engineering literature.

The allotriomorph provides two structural advantages for WF nucleation: it is already crystallographically related to the austenite on the opposite side of the grain boundary, providing a pre-existing lattice that satisfies the K-S relationship for plate growth; and it acts as a structural step or ledge from which individual WF sub-plates can propagate into the grain. The side-plates grow perpendicular or near-perpendicular to the grain boundary, fanning into the austenite grain in groups of 3–6 parallel plates known as a sheaf.

Interplate Regions: The Embrittling Component

The regions between growing WF plates are initially austenite that has been progressively enriched in carbon as the plates eject their interstitial content. As cooling continues through the bainite and martensite fields, this carbon-enriched austenite transforms to:

• Degenerate pearlite — at slow cooling rates, the carbon-enriched interplate austenite transforms to a coarse, unresolved pearlite with low interlamellar spacing. Relatively soft but brittle at low temperature.
• Martensite-austenite (M-A) constituent — at faster cooling rates typical of weld HAZ, the carbon-enriched interplate austenite partially transforms to hard, brittle martensite (with the remaining fraction as retained austenite). M-A islands are the single most damaging microstructural feature in WF-containing HAZ, with hardness values 50–200 HV above the surrounding WF plates, acting as nucleation sites for cleavage and ductile fracture.

Crystallography of Widmanstätten Ferrite

WF obeys the same crystallographic orientation relationships with the parent austenite as bainite and martensite, confirming the displacive component of its transformation mechanism. Two orientation relationships (OR) are observed:

Kurdjumov-Sachs (K-S) Orientation Relationship

K-S: {111}γ ∥ {110}α'   and   <1̅10>γ ∥ <1̅11>α

24 equivalent K-S variants per prior austenite grain.

Lath martensite, bainite, and WF all obey K-S (or near-K-S) OR —
confirming that all three involve a glissile (shear-capable) interface.

Nishiyama-Wassermann (N-W) Orientation Relationship

N-W: {111}γ ∥ {110}α'   and   <112>γ ∥ <110>α

N-W differs from K-S by 5.26° about the common {111}γ / {110}α normal.
12 N-W variants per austenite grain (vs 24 K-S).

WF habit planes:
  Low C (<0.2%):  near {111}γ
  Medium C (0.2–0.4%): near {112}γ
  (Higher C: WF field narrows and eventually disappears)

The crystallographic orientation relationship has a direct practical consequence: WF plates within a single sheaf are co-crystallographic, separated only by low-angle boundaries. A cleavage crack propagating along {100}α cleavage planes can traverse an entire sheaf of WF plates without encountering a high-angle grain boundary capable of deflecting or arresting it. The effective structural unit for cleavage fracture is therefore the WF sheaf, which can be 200–500 μm in dimension in a coarse prior austenite grain — far larger than the individual plate width and comparable to the prior austenite grain itself.

Effect of Prior Austenite Grain Size

Prior austenite grain size (PAGS) is the single most important microstructural variable controlling WF formation and severity. Its influence operates through three coupled mechanisms:

1. Nucleation site density — The austenite grain boundary area per unit volume scales as 2/d_γ where dγ is the mean grain diameter. A finer grain has more boundary area per unit volume, providing more nucleation sites for both allotriomorphic ferrite and WF. At very fine grain sizes (ASTM 8–10, d < 20 μm), the WF plates are physically limited in length by the grain diameter and the transformation produces very short, ineffective plates that contribute little to embrittlement.
2. Sheaf size limitation — Individual WF plates cannot exceed the prior austenite grain diameter in length. In coarse-grained austenite (ASTM 1–3, d = 200–500 μm), WF sheaves are large, producing the maximum cleavage path length and minimum impact toughness. In fine-grained steel (ASTM 7–9), WF plates are short and their toughness penalty is substantially reduced.
3. Tγs shift — Finer grain size slightly raises the effective nucleation temperature for allotriomorphic ferrite (more grain boundary area means faster grain-boundary ferrite coverage), reducing the amount of untransformed austenite available for WF growth below Tγs.

The relationship between grain size and toughness in the CGHAZ follows the well-established Hall-Petch type relationship for cleavage fracture, where the fracture stress σ_f is inversely proportional to the square root of the effective structural unit size. For WF-containing microstructures, the effective unit is the sheaf size rather than the individual lath width:

Hall-Petch for cleavage (Petch, 1954):
  σf = σi + ky × deff⁻½

where:
  deff = effective cleavage unit size (m)
         = sheaf size for WF, packet size for lath martensite

For WF-containing CGHAZ:
  deff typically 100–500 μm
  ⇒ σf low ⇒ low upper shelf energy, high DBTT

For acicular ferrite weld metal:
  deff typically 1–5 μm (random mutual impingement limits cleavage path)
  ⇒ σf high ⇒ high toughness

Widmanstätten Ferrite in Weld HAZ

The coarse-grained HAZ (CGHAZ) is the region of a steel weld where the base metal has been heated above approximately 1100–1200°C during the welding thermal cycle, causing austenite grain coarsening to ASTM grain size 1–4 (d = 100–500 μm). This coarse prior austenite grain structure, combined with the specific cooling rates that result from heat input and joint geometry, creates the ideal conditions for WF formation.

Heat Input and Δt_8/5

The cooling time between 800°C and 500°C, denoted Δt_8/5, is the primary process parameter controlling CGHAZ microstructure in steel welding. It is directly proportional to the arc energy (heat input) and inversely proportional to the preheat temperature and plate thickness for thin plates (2D heat flow) or plate thickness for thick plates (3D heat flow):

2D heat flow (thin plate: t < critical thickness):
  Δt8/5 = (4300 − 0.19 × T0²) × 10⁻³ × (Q / t²)

3D heat flow (thick plate: t > critical thickness):
  Δt8/5 = (6700 − 5 × T0) × 10⁻³ × Q

where:
  Δt8/5 = cooling time 800→500°C (seconds)
  T0  = preheat / interpass temperature (°C)
  Q    = arc energy = (V × I × η) / v  (kJ/mm)
         V = voltage, I = current, v = travel speed, η = efficiency
  t    = plate thickness (mm)

Typical Δt8/5 ranges and resulting CGHAZ microstructure:
  Δt8/5  <5 s:  martensite dominant
  Δt8/5  5–15 s:  lath martensite + bainite (low WF)
  Δt8/5  10–50 s: WF predominant + bainite or M-A
  Δt8/5  >50 s:  equiaxed F+P, coarse grain, low WF but poor toughness
               from grain coarsening alone

Effect on Impact Toughness: Quantitative Data

The 27 J Charpy criterion at −20°C specified by EN ISO 15614-1 for structural steel weld procedure qualification, and the 47 J at 0°C required by many offshore structural codes, are routinely at risk when CGHAZ contains more than 10–15 volume percent WF. Weld procedure qualification testing (WPQ) per ASME Section IX or EN ISO 15614 must include Charpy testing of the CGHAZ, with specimen location defined precisely to sample the highest WF content region.

Widmanstätten Ferrite vs Acicular Ferrite: The Toughness Contrast

Acicular ferrite (AF) and Widmanstätten ferrite are both plate-shaped ferrite products of austenite transformation, and both are visible in weld microstructures as needle-like features in optical micrographs. Their mechanical property consequences are, however, diametrically opposite. Understanding the structural reason for this contrast is central to weld metallurgy.

The toughness contrast arises primarily from the difference in high-angle boundary density. AF’s randomly oriented, mutually impinging plates create a very high density of high-angle (>15°) boundaries that deflect and arrest propagating cleavage cracks, consuming fracture energy at each boundary crossing. WF’s parallel sheaf structure creates essentially no high-angle boundaries within the sheaf, allowing long, uninterrupted cleavage crack propagation. The engineering goal in weld metal and CGHAZ is therefore to suppress grain-boundary nucleated WF and promote intragranular nucleated AF by controlling inclusion composition and distribution.

Suppression and Mitigation Strategies

Heat Input Control

Reducing welding heat input below approximately 1.5 kJ/mm increases the cooling rate through the WF formation window (Δt_8/5 < 10 seconds for typical structural steel thicknesses above 20 mm), suppressing WF in favour of bainite. This is the most direct and reliable control. For structural steels where minimum preheat is required for hydrogen cracking avoidance, the intersection of preheat requirements (which tend to slow cooling) and WF suppression (which requires faster cooling) defines a process window that must be explicitly managed in the weld procedure specification (WPS).

Titanium Microalloying and Grain Boundary Pinning

Titanium at 0.010–0.025 wt% combines with nitrogen (typically 60–100 ppm in modern steels) to form fine TiN precipitates (cube-shaped, <0.1 μm) during solidification. These precipitates pin austenite grain boundaries by the Zener pinning mechanism, preventing grain coarsening during the thermal cycle above 1100°C. The Zener grain size limit is:

Zener limiting grain diameter:
  dZ = (4r / 3fv) × (1/2)

where:
  r   = precipitate radius (m)
  fv  = precipitate volume fraction

For TiN in 0.015% Ti, 80 ppm N steel:
  fv ≈ 4.5 × 10⁻⁴,  r ≈ 25–50 nm
  dZ ≈ 75–150 µm (vs 200–500 µm without TiN pinning)

Result: finer PAGS limits WF plate length, reduces sheaf size,
and increases high-angle boundary density per unit area.
CVN improvement: typically +30 to +60 J at −40°C in CGHAZ
of Ti-treated vs untreated steel at identical heat input.

Post-Weld Normalising

Normalising at 880–920°C for 1 hour per 25 mm section thickness, followed by air cooling, eliminates WF by re-austenitising at a temperature where grain growth is limited (below the grain-coarsening temperature of ~1050°C for most carbon steels) and retransforming to equiaxed ferrite-pearlite on air cooling. The resulting ASTM grain size is typically 6–8 (d = 20–45 μm), which does not produce significant WF on subsequent cooling. Normalising is prescribed in several structural steel standards (EN 10025-3, API 2W) for plates that will be used in welded joints in low-temperature or dynamic loading service. The normalising heat treatment also relieves residual welding stresses, though it does not provide the same degree of stress relief as PWHT at lower temperature.

Multi-Pass Welding Procedure

In multi-pass welding (stringer bead technique), the deposition of successive weld beads reheats previously deposited metal and the underlying CGHAZ through an intercritical or supercritical thermal cycle. Regions of the first-pass CGHAZ that contained WF are re-austenitised at 800–1100°C by the heat of subsequent passes; on cooling, they transform to a refined equiaxed ferrite-pearlite or bainite microstructure depending on the cooling rate. This “bead-on-bead” refinement is the practical mechanism by which multi-pass welding with controlled interpass temperature consistently meets Charpy toughness requirements that single-pass or high-heat-input procedures cannot achieve.

Inclusion Engineering for Acicular Ferrite Promotion

Promoting intragranular nucleation of acicular ferrite in weld metal at the expense of grain-boundary nucleated WF requires controlling the weld metal inclusion population. Titanium-containing electrodes or wires produce a fine dispersion of TiO and TiN inclusions (0.1–0.5 μm diameter, 10⁷–10⁸/mm³) that are potent nucleation sites for AF. These inclusions nucleate AF preferentially because: they provide a low-energy incoherent interface for ferrite nucleation; their coefficient of thermal expansion mismatch with austenite creates local strain fields that reduce the nucleation energy barrier; and the MnS component of complex inclusions depletes the surrounding austenite in Mn, locally raising the Ae3 and promoting early ferrite nucleation. A detailed treatment of inclusion engineering is covered in the acicular ferrite and bainite guide.

Identification in Optical Micrographs

Reliable identification of WF requires systematic comparison with the other ferrite morphologies present in the same microstructure. The International Institute of Welding (IIW) ferrite morphology classification system (Buchmayr scheme) provides standardised criteria:

The key optical criteria for WF identification at 200–500× magnification in 2% nital etch are: (1) straight, sharply defined plate or lath boundaries; (2) parallel arrangement within groups (sheaves) of 3–8 plates; (3) relationship to prior austenite grain boundary (plates emanate from GB or allotriomorph layer); (4) coarse dark interplate regions identifiable as M-A or pearlite; and (5) plate width 5–20 μm (much coarser than bainite). Confirmation of the crystallographic relationship requires EBSD (electron backscatter diffraction) for orientation mapping.

Industrial Case Studies

Offshore Jacket Weld HAZ Toughness Failure

During weld procedure qualification testing for a North Sea offshore jacket (S355G10+M steel, 50 mm plate), Charpy specimens extracted from the CGHAZ of a submerged arc weld (SAW) at heat input 6.5 kJ/mm showed a mean CVN of 18 J at −20°C against a minimum requirement of 27 J. Metallographic examination revealed approximately 35 volume percent WF with large M-A islands (average 3.5 μm) in the CGHAZ. Corrective actions: heat input reduced to 3.2 kJ/mm using a tandem-wire SAW procedure with controlled interpass temperature ≤200°C; WF content reduced to <8 vol%; re-tested CVN mean of 62 J at −20°C.

Cast Steel Structural Node

A cast carbon steel structural node (0.20C-1.2Mn, normalised condition) for a truss bridge exhibited brittle fracture during proof loading at −15°C. Fractographic analysis showed cleavage fracture initiating from the casting skin region. Metallographic cross-section of the skin zone revealed coarse WF with 300–400 μm PAGS and approximately 25% M-A content — a consequence of the slow cooling rate of the as-cast skin during solidification, which produced WF despite subsequent normalising (the normalising had been performed at 850°C but the grain-coarsening effects of the original solidification were not fully eliminated at this normalising temperature). Remedial action: re-normalising at 900°C for 3 hours eliminated WF; CVN improved from 12 J to 78 J at −20°C.

Frequently Asked Questions

Recommended References

Disclosure: MetallurgyZone participates in the Amazon Associates programme. If you purchase through these links, we may earn a small commission at no extra cost to you. This helps support free technical content on this site.

Further Reading