Intermetallic Phases — What Engineers Must Know

Intermetallic phases are among the most consequential microstructural features an engineer will encounter in fabricated components. Unlike simple solid solutions, these ordered compounds — sigma, Laves, chi, mu, and others — form at discrete compositions, carry their own crystal structures, and almost always degrade the properties of the host alloy. Understanding when they form, why they form, and how to prevent or remove them is not academic; it is a daily concern in duplex stainless fabrication, Cr-Mo power piping, nickel superalloy repair welding, and any service environment that combines high temperature with aggressive chemistry.

In welding practice, intermetallics typically precipitate in two locations: the heat-affected zone (HAZ) of the base metal, where the thermal cycle holds the material within a critical temperature window long enough for nucleation and growth; and the weld metal itself, where composition imbalances or slow post-weld cooling promote the same reactions. The consequences range from catastrophic brittle fracture at ambient temperature to accelerated pitting corrosion and stress corrosion cracking. This guide covers the principal intermetallic phases encountered in industrial fabrication, their metallurgical origins, detection methods, relevant ASTM and ASME code implications, and practical prevention strategies.

Whether you are qualifying a welding procedure for super duplex piping, specifying heat treatment for a P91 pressure vessel, or troubleshooting a failed Charpy test on austenitic stainless weld coupons, this article gives you the fundamental knowledge to diagnose and resolve intermetallic-related problems with confidence. Topics such as delta ferrite control and duplex stainless welding are tightly coupled to intermetallic phase management and are cross-referenced throughout.

1. What Is an Intermetallic Phase?

The term “intermetallic” describes an ordered, stoichiometric or near-stoichiometric compound formed between two or more metallic elements whose crystal structure differs from those of the constituent metals. The classic definition, attributed to G. E. R. Schulze, covers solid phases with two or more metallic elements whose atomic positions are filled in a regular, repeating pattern — distinct from the random substitutional arrangement in a solid solution. The result is a structure that behaves more like a ceramic than a metal: hard, brittle, resistant to dislocation motion, and often highly ordered right up to its melting point.

In the context of Fe-Cr-Ni-Mo engineering alloys, most problematic intermetallics belong to the family of topologically close-packed (TCP) phases. This group includes sigma, Laves, chi, mu, and R-phase. They share a common characteristic: their crystal structures are built from close-packed layers of atoms with alternating large and small atomic radii, creating a rigid framework that inhibits plastic deformation and acts as a stress concentrator under load.

Why Intermetallics Form

Intermetallics form because certain alloy compositions minimise their Gibbs free energy by segregating elements into an ordered compound rather than keeping them in solid solution. The driving forces are:

• Composition: High concentrations of Cr, Mo, W, Nb, or Si relative to Fe and Ni push the alloy into a two-phase or multi-phase field on the phase diagram.
• Temperature: Diffusion must be fast enough to allow atoms to redistribute, yet the temperature must be below the phase’s solvus. This creates a critical window, typically 600°C to 950°C for sigma in stainless steels.
• Time: At very short hold times, nucleation may not occur even at the optimal temperature. Intermetallics in duplex stainless can appear in as little as a few minutes at peak kinetic temperatures; in fully austenitic grades, hours may be required.
• Prior microstructure: Ferrite transforms far more readily to sigma than austenite does. The bcc lattice of ferrite is crystallographically closer to the sigma structure, and diffusion coefficients in ferrite are orders of magnitude higher than in austenite at the same temperature.

2. Sigma Phase (σ)

Sigma phase is the most widely documented and commercially significant intermetallic in austenitic and duplex stainless steels. It is a topologically close-packed compound based primarily on the Fe-Cr binary system, with a primitive tetragonal unit cell containing 30 atoms. Its composition in practical alloys ranges roughly from FeCr to Fe_xCr_yMo_z, with molybdenum strongly promoting its formation by raising the electron-to-atom ratio into the favourable range (approximately 6.2 to 7.0).

2.1 Composition and Crystal Structure

The sigma phase unit cell contains 30 atoms arranged in five crystallographically distinct sites (A through E), each preferentially occupied by atoms of a particular size. In Fe-Cr alloys, Cr typically occupies two of the five sites; iron fills the remaining three. The addition of Mo, Mn, Ni, or Si redistributes this site occupancy and shifts the formation kinetics. The resulting compound is non-magnetic, with a Vickers hardness of 900 to 1,000 HV and a density close to that of the matrix — making it difficult to detect by simple visual or magnetic inspection.

2.2 Formation in Duplex Stainless Steels

In standard 2205 (UNS S31803 / S32205) and super duplex 2507 (UNS S32750) grades, sigma nucleates preferentially at ferrite-austenite boundaries and ferrite-ferrite-ferrite triple junctions. The driving mechanism is the decomposition of delta ferrite: Fe, Cr, and Mo diffuse rapidly through the bcc ferrite lattice, allowing sigma and secondary austenite (γ₂) to form simultaneously. The overall reaction can be represented schematically as:

In multi-pass welding, the fill passes reheat the root and intermediate beads into the sigma-forming window. This is the single most common cause of ASTM A923 Method B or Method C failures in duplex weld procedure qualifications. Controlling interpass temperature to below 150 °C and keeping heat input within the 0.5 to 2.5 kJ/mm range specified in most client welding procedure specifications is essential.

2.3 Formation in Austenitic Stainless Steels

Fully austenitic grades (304, 316, 321, 347) are significantly more resistant to sigma formation than duplex grades because diffusion in the fcc austenite lattice is much slower than in bcc ferrite. However, grades containing delta ferrite in the weld metal — which is deliberately introduced to prevent solidification cracking in austenitic welds — provide rapid diffusion pathways that promote sigma after long service exposures at 600°C to 900°C. High-Mo grades (316L, 317L) are particularly susceptible. In the stainless weld decay mechanism, sensitisation to chromium carbide precipitation is distinct from sigmatisation but can occur simultaneously. See the dedicated article on stainless steel weld decay for the full intergranular corrosion picture.

3. Laves Phase (η)

Laves phases are a family of intermetallic compounds with the general formula AB₂, where B atoms are approximately 1.225 times smaller than A atoms. This size ratio allows particularly efficient close-packing. The three structural polytypes are C14 (hexagonal, MgZn₂-type), C15 (cubic, MgCu₂-type), and C36 (dihexagonal). In engineering alloys, the most practically important Laves phases are:

• Fe₂Mo and Fe₂Nb in high-Mo ferritic and austenitic stainless steels
• Fe₂W and Cr₂(W,Mo) in P91 and P92 creep-resistant Cr-Mo steels
• Fe₂Ti in Ti-stabilised stainless grades
• Ni₂Nb in Inconel 625 and alloy 718

3.1 Laves Phase in P91 and P92

The Cr-Mo steels P91 (9Cr-1Mo-V-Nb) and P92 (9Cr-0.5Mo-1.8W-V-Nb) derive much of their creep strength from M₂₃C₆ carbides and MX-type fine precipitates (V and Nb carbonitrides). During long-term service above 550°C, these fine strengthening precipitates coarsen while a coarse Laves phase — primarily Fe₂(W,Mo) in P92 — nucleates at lath boundaries and prior austenite grain boundaries. This Laves phase consumes W and Mo from the matrix, reducing the solid-solution strengthening contribution and accelerating the coarsening of MX precipitates. The combined effect is a progressive reduction in creep rupture strength. For a comprehensive treatment of P91 welding and metallurgy, see the P91 welding guide on this site.

3.2 Laves Phase in High-Si and High-Mo Ferritic Stainless

Ferritic grades such as Type 430 and 441 can form Fe₂Nb or Fe₂Ti Laves phase during slow cooling through 500°C to 800°C. In thin sections welded without filler, the weld metal and HAZ may contain enough Laves phase to show a 475°C embrittlement-type toughness loss compounded by the Laves effect. Solution annealing above 950°C followed by water quench is the standard remedy.

4. Chi Phase (χ)

Chi phase is a body-centred cubic (bcc) intermetallic, structurally related to the alpha-manganese structure, with up to 58 atoms per unit cell. In Fe-Cr-Mo alloys, chi phase has a nominal composition of approximately Fe₃₆Cr₁₂Mo₁₀, making it significantly enriched in molybdenum relative to sigma. The higher Mo content means chi phase is particularly prevalent in super duplex (2507) and supermartensitic stainless steels, as well as high-Mo austenitic grades.

4.1 Relationship Between Chi and Sigma

Chi phase often precipitates earlier than sigma at a given temperature because its composition is closer to that of the ferrite phase from which it nucleates. Once chi has precipitated, it can dissolve as sigma grows and consumes the Mo and Cr that stabilised chi. In prolonged thermal exposures, the final microstructure at 850°C is dominated by sigma rather than chi. This transition is diffusion-controlled: chi is the kinetically favoured product at short times; sigma is the thermodynamically stable endpoint. Both phases are equally detrimental to impact toughness and corrosion resistance during the period when they coexist.

5. Mu Phase (μ) and R-Phase

5.1 Mu Phase

Mu phase (also written μ-phase) has an ideal stoichiometry of A₆B₇ with the prototype structure W₆Fe₇, forming a rhombohedral unit cell with 13 atoms. In practical alloys, mu phase carries high concentrations of W, Mo, and sometimes Re, with Fe and Co filling the remaining positions. It is most commonly encountered in nickel-base superalloys and high-W steels such as P92 during long-term service above 600°C. Mu phase reduces creep ductility by consuming tungsten and molybdenum from the matrix, similar in effect to Laves phase in P91/P92. In some CrNiMo stainless steels, mu phase forms alongside sigma during prolonged exposure, accelerating grain boundary embrittlement.

5.2 R-Phase

R-phase is a hexagonal TCP compound with a large unit cell (159 atoms) and a composition close to Mo₆Fe₇Cr in austenitic stainless steels. It forms at slightly lower temperatures than sigma — typically 550°C to 750°C — and is favoured in high-Mo grades (317L, 904L, 254 SMO) after prolonged service or slow cooling. R-phase is generally considered less detrimental than sigma or chi because it is less brittle, but it still depletes Mo from the matrix and can serve as a nucleation site for the more harmful sigma phase during subsequent thermal exposure.

6. Intermetallic Phases in Nickel-Base Alloys

Nickel-base alloys occupy a unique position: they rely on intermetallic precipitates for their mechanical strength, yet are simultaneously vulnerable to detrimental TCP phases if composition or thermal management is incorrect. The distinction between beneficial and harmful intermetallics is fundamental to alloy design in this family.

6.1 Beneficial Intermetallics: Gamma-Prime and Gamma-Double-Prime

Gamma-prime (γ′, Ni₃Al or Ni₃(Al,Ti)) is the primary strengthening precipitate in wrought superalloys such as Inconel 718 and Waspaloy. It forms as coherent, ordered L1₂ precipitates within the austenite matrix, impeding dislocation glide and providing the high-temperature strength essential in gas turbine applications. Gamma-double-prime (γ″, Ni₃Nb with D0₂₂ structure) performs the same function in Nb-rich alloys like Inconel 718. These precipitates are intentional; maintaining them requires strict control over ageing heat treatment temperatures.

6.2 Detrimental TCP Phases in Ni Alloys

When the alloy composition deviates from design limits — through dilution with dissimilar filler metals, segregation during weld solidification, or composition variability between heats — TCP phases precipitate at grain boundaries and interdendritic regions. The most common detrimental intermetallics in Ni-alloy welds are:

• Laves phase (Ni₂Nb) in Inconel 625 and 718: forms during weld solidification due to Nb segregation to the final liquid. Reduces ductility and rupture strength. Minimised by limiting Nb segregation through low heat input and post-weld solution annealing above 1,150°C.
• Sigma and mu phases in high-Mo, high-W nickel alloys: form during prolonged service at 700°C to 900°C. Detected by SEM-EDS and XRD on extracted residues.
• Delta phase (δ, orthorhombic Ni₃Nb) in Inconel 718: forms above approximately 900°C during overageing, consuming Nb that would otherwise contribute to strengthening γ″ precipitates.

7. Comparison of Principal Intermetallic Phases

8. Effects on Mechanical and Corrosion Properties

8.1 Impact Toughness

The most immediate engineering consequence of TCP phase formation is a catastrophic loss of Charpy impact toughness. Sigma phase at even 5 vol.% in a duplex stainless steel can reduce the upper-shelf Charpy energy from a typical 200 to 250 J (in the solution-annealed condition) to near-zero values at ambient temperature. This is because sigma is both hard (HV ~ 950) and brittle, with no slip systems available at room temperature. Cracks initiate at sigma particles and propagate across grain boundaries with minimal plastic work. The transition to brittle behaviour is essentially instantaneous: a single multi-pass weld that passes through the sigma window multiple times may produce a Charpy failure on a 10 × 10 mm specimen even when the sigma volume fraction is too low to detect by optical microscopy.

8.2 Corrosion Resistance

Sigma, chi, and mu phases are all enriched in Cr and Mo relative to the alloy’s nominal composition. Their precipitation removes these elements from the surrounding matrix, creating Cr- and Mo-depleted zones that are highly susceptible to localised corrosion. In duplex stainless steels, a reduction in critical pitting temperature (CPT) of 20°C to 30°C is typical for sigma fractions of 1 to 3 vol.%. In service environments meeting the criteria of sour service (H₂S environments per NACE MR0175 / ISO 15156), intermetallic phases also increase susceptibility to sulphide stress cracking and hydrogen-induced cracking by providing high-hardness crack initiation sites. Corrosion testing to ASTM G48 Method A or F (elevated-temperature ferric chloride test) is routinely used to screen for intermetallic-induced pitting. The methodology is described in the ASTM G48 corrosion testing guide.

8.3 Creep and High-Temperature Properties

In Cr-Mo and Ni-base alloys intended for elevated-temperature service, TCP phase formation during welding or PWHT directly reduces the long-term creep rupture strength. In P92 steel, excessive Laves phase precipitation reduces the 100,000-hour rupture strength at 600°C by up to 25% compared with the design curve values. The mechanism is a combination of solid-solution depletion (W and Mo removed from the matrix) and the pinning of recovery processes: the coarse Laves particles impede subgrain boundary migration initially but do not prevent it during very long exposures, accelerating creep damage accumulation at these locations.

9. Prevention Strategies During Welding

9.1 Duplex and Super Duplex Stainless Steels

• Heat input control: Hold heat input between 0.5 and 2.5 kJ/mm. The lower bound prevents excessive ferrite in the weld metal; the upper bound prevents excessive time in the sigma formation window.
• Interpass temperature: Limit to 150°C maximum. Use a contact pyrometer on the weld cap, not an infrared gun, for reliable measurement in shielded environments.
• Shielding gas: Use 98% Ar + 2% N₂ for TIG (GTAW) root and fill passes. Nitrogen additions suppress secondary austenite formation and help maintain the target 40 to 60% ferrite balance.
• Back purging: For root passes in pipe, back purge with pure Ar or 90% Ar + 10% N₂ to prevent nitrogen loss and oxidation. See the GTAW welding guide for back-purge setup details.
• Avoid PWHT in the sigma window: Stress relief is generally not applied to duplex stainless; if dimensional stabilisation is required, use mechanical stress relief or vibration stress relief below 300°C.

9.2 Cr-Mo Steels (P91, P92)

• Control PWHT temperature strictly within the specified range (730°C to 780°C for P91). Use calibrated Type K thermocouples within 50 mm of the weld centreline.
• Avoid slow cooling between 450°C and 300°C during post-weld cooling; this window promotes temper embrittlement through Laves and M₂X coarsening at grain boundaries.
• Filler metal selection: match composition to the base metal specification (ER90S-B9 / E9015-B9 for P91; ER90S-G / E9015-B9 modified for P92). Over-alloyed fillers increase the risk of TCP phases in the dilution zone.

9.3 Nickel-Base Alloys

• Keep heat input low to minimise Nb segregation to the interdendritic regions in 625 and 718. Travel speed should be on the higher end of the qualified range.
• For Inconel 625 weld cladding and buttering, specify a minimum 1,150°C solution anneal after welding if the component geometry permits, to dissolve any Laves phase formed during solidification.
• Avoid overheating the HAZ of age-hardened superalloys above the γ″ or γ′ solvus, as this dissolves the strengthening precipitates and promotes TCP formation on re-ageing.

10. Detection and Testing Methods

10.1 Non-Destructive Indicators

No single non-destructive method definitively identifies intermetallic phases in a finished weld, but the following provide reliable screening:

• Ferrite number measurement (Feritscope or Fischer FMP30): A significant drop in FN below the expected post-weld value indicates ferrite decomposition and is consistent with sigma formation. Calibrate per AWS A4.2 for duplex grades.
• Hardness survey (HV10 Vickers): Localised hardness spikes above HV 350 in duplex stainless, or above HV 265 in P91 welds, indicate possible TCP phases or inadequate PWHT.
• Impact testing (ASME / ASTM A923 Method B): Charpy impact testing at −40°C per ASTM A923 is one of the most sensitive indicators of sigma in duplex stainless. A minimum absorbed energy of 54 J is typically required; values below 27 J are an immediate rejection criterion.

10.2 Destructive and Analytical Methods

11. Remediation — Dissolving Intermetallic Phases

Once intermetallic phases have formed, the remediation options are limited but well-defined:

11.1 Solution Annealing

For duplex and austenitic stainless steels, solution annealing at 1,050°C to 1,100°C (for standard 2205) or 1,080°C to 1,150°C (for super duplex 2507) is the standard remedy. Hold time depends on section thickness: a minimum of 30 minutes per 25 mm of section, followed by water quenching. The quench rate through the 600°C to 950°C window must be rapid enough to prevent re-precipitation; air cooling is generally insufficient for sections above 10 mm in super duplex.

11.2 Re-tempering of P91/P92

For P91 where PWHT was performed incorrectly, re-tempering at the correct temperature (730°C to 780°C) after a new normalising treatment (if accessible) may partially restore properties. However, if sigma or Laves phase has coarsened significantly, full normalise-and-temper is the only reliable approach. On finished assemblies, this is typically not feasible, requiring weld removal and replacement.

11.3 Weld Removal and Repair

Where geometry does not permit solution annealing, the affected weld must be removed by arc gouging or mechanical grinding to sound base material, and re-welded under the corrected procedure with verification that interpass temperatures and heat inputs comply with the WPS. Post-repair testing per the original qualification standard (ASTM A923, mechanical testing) confirms restoration of properties.

12. ASME and AWS Code Implications

Intermetallic phase formation is not directly addressed by name in ASME Section IX, but its consequences — toughness reduction and corrosion attack — trigger requirements in several code sections:

• ASME Section VIII Division 1, UG-84 and UHA-51: Charpy impact testing requirements for high-alloy steels at design temperatures. A weld displaying sigma-induced toughness loss will fail the minimum impact values required by Table UHA-51. See the detailed guide to UG-84 impact testing.
• ASME B31.3 Chapter IX (High-Purity Piping): Duplex stainless piping in high-purity service is frequently tested against ASTM A923; compliance is referenced in B31.3 or client specifications.
• AWS D1.6 (Structural Stainless Steel Welding): Procedure qualification for duplex stainless includes requirements for ferrite measurement and Charpy testing that effectively screen for sigma phase.
• NACE MR0175 / ISO 15156: Duplex stainless in sour service must have ferrite content within the specified range; excessive sigma formation (which alters the ferrite reading) is therefore a disqualifying condition for wet H₂S service.
• P-Number and F-Number assignments: Duplex stainless steels are assigned P-Number 10H in ASME Section IX, with filler metals matched accordingly. Deviations from qualified base metal or filler composition can push the dilution zone into the TCP formation field. Understand your P-Number and F-Number assignments before selecting consumables.

Recommended Reference Books

Frequently Asked Questions

Further Reading on WeldFabWorld