3-4. Hydrogen induced stress cracking resistance of precipitation hardened nickel-based alloys using the slow strain rate tensile test method

I.Salvatori, L. Alleva,

Rina Consulting Centro Sviluppo Materiali , Italy

Abstract: Precipitation Hardened (PH) Ni-based alloys have proved to be sensitive to Hydrogen Induced Stress Cracking (HISC), and HISC related failures in the Oil and Gas industry have been experienced in the case of UNS N07718, UNS N07725 and UNS N07716. Slow Strain Rate Tensile (SSRT) tests conducted under cathodic polarization gave encouraging results as a mean to evaluate HISC resistance when applied to UNS N07718, enabling the discrimination of acceptable and unacceptable microstructures as according to API 6A CRA.

As a consequence, an extensive test program, within the framework of a Joint Industrial Project sponsored by several petroleum companies, was launched to verify the validity of this method on several PH Ni grades.

The behavior of several Precipitation Hardened (PH) Ni-based alloys with respect to Hydrogen Induced Stress Cracking resistance was studied using the Slow Strain Rate Tensile test method carried out under hydrogen charging conditions (applying a constant cathodic current density throughout the test).

The main objectives of this program were to develop a test method to allow for the evaluation of HISC resistance, in order to rank materials and possibly define acceptance criteria for each material, and also, to better understand the relationship between microstructure and HISC resistance.

Several industrial heats of UNS N07718, UNS N09945/945X, UNS N09925, UNS N09935, UNS N07725 and UNS N07716 were evaluated. A detailed microstructural analysis was performed on each heat, at different levels, involving SEM and TEM examinations at high magnification on etched samples, to reveal the phases present at grain boundaries.

Deep microstructural investigations were carried out to characterize the different alloys, focusing on the precipitates at grain boundaries: nature of precipitates, morphology, distribution, grain boundary coverage.

A methodology was developed to quantify the grain boundary coverage (defined as the length of precipitates at grain boundary divided by the total grain boundary length), improving upon the visual evaluation method proposed in the standard. Other features, such as grain size, level of primary carbides, as well as alloy strength, were also considered, to rank the materials with respect to their hydrogen embrittlement resistance as evaluated by SSRT tests under cathodic protection. Elemental chemical analysis of the precipitates was also conducted by Energy Dispersive X-ray spectrometry.

The performance of the different alloys with respect to HISC resistance was discussed on the basis of the plastic elongation obtained in the SSRT tests under cathodic polarization, the microstructural features and the microstructural criteria given in the API 6A CRA standard.

Before starting the analyses on the selected alloys, a deep fine tuning of the procedure was carried out in order to find and define the optimized experimental parameters to be used.

The following parameters were fine-tuned:

• Magnification needed to have a good resolution;

• Number of fields to evaluate, to cover as far as possible a representative area of the specimen;

• Evaluation of method to use to count grain boundary precipitates (counting, comparison, etc....);

• Microstructural parameters to be analysed.

Image analysis was performed on about 50 micrographs captured at high magnification (x5000 to x10000), to characterize the precipitates (equivalent diameter, area, aspect ratio).

The grain boundary coverage was measured, using the number of particles located at the grain boundary and the length of grain boundary on the analyzed image. For some heats, Transmission Electron Microscopy (TEM) investigations were also carried out.

The mechanical properties of the different materials tested were quite homogeneous for three different sampling locations and complied with the requirements of the API 6ACRA standard. With respect to the micrographic evaluation, most of the heats displayed acceptable microstructures, However, in a few cases the microstructure was identified as ‘mixed’, since in the same sample both acceptable and unacceptable zones could be identified.

In all the specimens, the precipitates displayed two main morphologies, small blocks and platelets. For all the analysed Ni alloys grades, the blocky coarse precipitates within the gamma grains were Nb and Ti carbides. Some carbides rich in Nb, Mo and Ti were also found at grain boundaries. Thinner round shaped or elongated precipitates showed a higher Cr concentration (CrMo₂₃C₆).

Very small precipitates at grain boundaries, as well as intergranular (, , ) and intragranular (’ and ’’) phases were detected by TEM analysis. The chemical analyses conducted on these precipitates indicated that carbides as well as  phase, ,  phases could be present at the grain boundary.

No obvious correlation between the quantified microstructural parameters and the ductility of the alloys tested with SSRT under cathodic protection was found. It seems that no single parameter is responsible for the hydrogen embrittlement susceptibility and not all of the parameters have the same role for the different alloy families.

The grain boundary coverage parameter, as recommended in API 6ACRA as a primary controlling feature, was not found to correlate uniquely to the results for the various alloys; although it was important for UNS N07718. Furthermore, with the same percentage of grain boundaries coverage, the various alloys can have very different ductility. The reason is that the nature of precipitates cannot be ignored.

The primary carbides volume fraction did not match the plastic elongation values in any clear way, suggesting that this microstructure feature is not so important for hydrogen embrittlement resistance of Ni alloys. Nevertheless, an intense study was carried out to achieve information from the fracture surfaces of SSRT specimens. The fracture surface propagation mode clearly indicates that primary carbides play a role, especially at an early stage of fracture propagation. They can be traps for hydrogen. Around them hydrogen can accumulate and weaken the atomic bonds. Voids form (vacancies) and fracture initiate and propagates via void coalescence.

An interesting correlation is the one regarding the grain size. All the figures indicate that as the grain size increases the hydrogen embrittlement resistance improves. This can be a further confirmation that hydrogen embrittlement phenomena are activated at grain boundaries; large grains imply lower extension of the boundaries and high energy required by dislocations to transport hydrogen (high mean free path). The strain (dislocation movement) is responsible for hydrogen accumulation at specific sites as precipitates at grain boundaries. Microvoids form when hydrogen supersaturates the traps. Fracture surface analysis revealed, in some cases, that this mechanism also occurred at twinning sites.

The effect of the alloy mechanical resistance is well known and was largely confirmed by this project.

Considering all these aspects it can be confirmed that a simple evaluation of precipitates at grain boundaries, as proposed by the API 6A CRA, is not sufficient to indicate the hydrogen embrittlement susceptibility of the different PH Ni alloys included in the API standard.