Carbide Precipitation and Temper Embrittlement in CrMo Alloy Steel Pipe Fittings: PWHT Windows and Cooling Rate Control
CrMo alloy steel pipe fittings (P11, P22, P5, P91, P92) undergo post-weld heat treatment (PWHT) to temper the martensite or bainite microstructure formed during welding, reduce residual stress, and restore toughness in the weld HAZ. However, if PWHT is performed incorrectly — particularly if the cooling rate after PWHT is too slow, or if the PWHT temperature is outside the correct range — two distinct embrittlement mechanisms can occur: temper embrittlement from trace element segregation, and carbide precipitation from over-tempering or re-heating in the incorrect temperature window. Both mechanisms reduce toughness in a way that may not be apparent from post-PWHT hardness testing but becomes critical under low-temperature or impact loading conditions.
Temper Embrittlement: Mechanism and Temperature Range
Temper embrittlement in CrMo steels occurs when the material is held or cooled slowly through the embrittlement range of approximately 370–560°C. During this temperature exposure, trace impurity elements — primarily phosphorus (P), antimony (Sb), arsenic (As), and tin (Sn) — segregate from the bulk to grain boundaries. At grain boundaries, these elements displace carbon and reduce the grain boundary cohesive strength, making the material susceptible to brittle intergranular fracture at impact loading or low temperature. The embrittlement is not apparent from room-temperature tensile testing or hardness measurement — the yield strength and UTS are unchanged. It is only revealed by Charpy V-notch impact testing at low temperature (typically −20°C or −30°C) or by the shift in ductile-to-brittle transition temperature (DBTT). Temper embrittlement is reversible: re-heating above 600°C for an adequate time re-dissolves the grain boundary segregates back into the bulk, restoring full toughness. However, if the material is cooled slowly again through 370–560°C, the segregates re-accumulate and embrittlement returns. This means that a CrMo fitting that was correctly PWHT'd and initially met Charpy requirements can become embrittled by a subsequent slow cooling event — such as an unplanned furnace cooling during a maintenance shutdown or a repeated PWHT cycle with inadequate cooling rate control.
Controlling PWHT Cooling Rate
To avoid temper embrittlement, the cooling rate from PWHT must be fast enough to pass through the 370–560°C embrittlement range without allowing significant impurity segregation. For most CrMo grades (P11, P22, P5, P9), a cooling rate of at least 55°C/hour (approximately 1°C/minute) from the PWHT temperature to below 370°C is sufficient to prevent significant embrittlement. For large-section fittings where the centre cools slower than the surface, forced air cooling or water spray may be required to maintain adequate cooling rate through the embrittlement range. This contrasts with the common fabrication practice of allowing fittings to cool in the furnace after PWHT (furnace cool), which typically produces cooling rates of 10–30°C/hour through the embrittlement range — well below the threshold. For P91 and P92 (grade-specific concern), the PWHT cooling rate is even more critical because the Ac₁ temperature (the lower critical transformation temperature at which austenite starts to form on heating) for P91 is approximately 820°C. If PWHT is carried out above Ac₁ even briefly, partial re-austenitisation occurs, and the subsequent cooling creates fresh untempered martensite — a significantly harder and more brittle microstructure than the properly tempered martensitic structure. P91 PWHT must be performed below Ac₁ (typically 730–780°C), and this temperature window is narrower than for P22 (PWHT at 690–750°C, Ac₁ approximately 810°C).
Step-Cooling Test for Temper Embrittlement Susceptibility
For critical P22 fittings — high-energy steam lines, nuclear class, and refineries with a history of temper embrittlement failures — the material's susceptibility to temper embrittlement is characterised by the step-cooling test (ASTM A540 Annex or ASME Code Case). In this test, a set of Charpy specimens is given a controlled heat treatment that includes a slow-step cooling through the embrittlement range (the "step-cool" cycle), and the Charpy impact energy after step-cooling is compared to the as-PWHT'd value. The increase in DBTT (ΔT = transition temperature after step-cool minus transition temperature as-PWHT'd) characterises the material's susceptibility. Low susceptibility: ΔT < 15°C. High susceptibility: ΔT > 40°C. High susceptibility indicates high impurity content in the heat — the J-factor (J = (Mn + Si)(P + Sn) × 10⁴) and X-factor (X = (10P + 5Sb + 4Sn + As)/100) are calculated from the chemistry to predict susceptibility. Purchase orders for P22 fittings in high-energy power plant service commonly specify maximum J-factor (≤ 180) and maximum X-factor (≤ 15) as additional chemistry acceptance criteria beyond the standard ASTM A234 WP22 requirements.
Carbide Over-Precipitation in P91
In P91 (9Cr-1Mo-V-Nb), the tempering reaction during PWHT involves precipitation and coarsening of M₂₃C₆ carbides and MX (VN, NbC) precipitates. If PWHT temperature is at the high end of the specification (above approximately 790°C) or if PWHT is held for too long, M₂₃C₆ carbides coarsen excessively — the precipitate strengthening contribution decreases, the solid solution hardening from Mo is reduced, and the creep strength drops below the ASME allowable values used in the pipe stress design. This is the same microstructural mechanism that causes in-service creep degradation (see the creep damage article) but occurring during PWHT rather than in service. For P91 fittings, PWHT time-temperature records must be reviewed carefully — both too-short/too-low (insufficient tempering, high residual stress, brittle HAZ martensite) and too-long/too-high (over-tempered, reduced creep strength) are non-conforming conditions that require engineering disposition before the fitting is accepted for service.