Hydrogen permeation in 13% Crsuper martensitic stainless steel and API X70 pipeline steel

Abstract

Hydrogen permeation measurements have been carried out on 13% Cr super martensitic stainless steel (SMSS), API X70 steel and type HBQ Coreweld under cathodic charging conditions at temperatures from 5°C to 80°C. In addition, testing of SMSS was performed at hydrostatic pressures from atmospheric pressure to 100 bar, in artificial seawater at 5°C and 20°C. Moreover, hydrogen diffusion was measured in SMSS at 0% strain and at 0.4% strain (yield strength). The thickness of the specimens was 300 μm. The bulk diffusion control was verified by comparing the subsequent permeation transient and assuring that these transients have the same shape and steepness when plotted in dimensionless parameters.

New equipment has been developed to study hydrogen diffusion in metals according to ISO 17081 [1] at different temperatures, mechanical loads and hydrostatic pressures. The principle of the equipment was based on the Devanathan-Stachurski electrochemical permeation technique where hydrogen is produced electrochemically in one cell, diffuses through the specimen and is detected in the other cell. The equipment consists of two titanium autoclaves integrated with a loading unit. The loading unit allows testing at constant load, fatigue load and slow strain-rate. The temperature can vary from 4°C to 80°C using a cooling/heating system. The autoclaves can be pressurized at the maximum pressure of 100 bar by using gas supplied from an external reservoir.

The effective hydrogen diffusion coefficient for SMSS at room temperature was in the range 1.7 ・ 10−9 to 4.0 ・ 10−9 cm2 s−1. The sub-surface hydrogen concentration at room temperature was around 5-10 ppm.

Hydrogen diffusivity in SMSS appeared to follow Arrhenius equation and according to Deff = 10.61 ・ 10−2 exp (−42.8 kJ mol−1/RT). The experimental transients were steeper than predicted by the solution of Fick’s 2nd law which indicated significant reversible trapping with high fractional occupancy of trap sites. The similarity of the first and the second permeation transients indicated that irreversible trapping was insignificant in this steel.

The retained austenite content in as-received SMSS test material was measured as 19%. After 0.4% strain this value appeared to be around 3%. This decrease in the retained austenite was attributed to deformation-induced martensitic transformation.

The binding energy of reversible traps in SMSS was calculated to be 36.4 kJ mol−1 which was consistent with literature. The reversible trap density was measured as 6.8 ・ 1021 sites cm−3. After 0.4% strain the binding energy remained unchanged while the traps density decreased to 3.5 ・ 1021 sites cm−3. This was probably due to the decrease in the retained austenite content after 0.4% deformation and the decreased length of martensite/austenite grain boundaries that act as reversible trap sites. Effective hydrogen diffusion coefficient increased moderately after 0.4% strain as compared to as-received material. This was also attributed to deformation-induced martensitic transformation and faster diffusivity of hydrogen in martensite than in austenite.

Hydrogen absorption in SMSS appeared to be unaffected by 100 bar hydrostatic pressure as compared with testing at atmospheric pressure. Hydrogen permeation was measured by two methods and neither method showed any impact of 100 bar hydrostatic pressure on hydrogen permeation. This is attributed to an increase in hydrogen fugacity, i.e. the formation of hydrogen gas, at increased pressures.

Calcareous deposits formed at 5°C and at 20°C in artificial sea water did not have any significant effect on hydrogen permeation current density.

For API X70 steel grade and for type HBQ weld hydrogen diffusion was according to the Arrhenius equation. Deff at room temperature for X70 base metal and HBQ weld was 7.95 ・ 10−7 cm2 s−1 and 1.3 ・ 10−6 cm2 s−1, respectively. The relationship between Deff and temperature was Deff = 13.87 ・ exp (−41.3 kJ mol−1/RT) and Deff = 7.33 ・ exp (−38.4 kJ mol−1/RT) for X70 steel and HBQ weld respectively. The experimental transients were steeper than predicted by Fick’s law which indicated significant reversible trapping with high fractional occupancy of trap sites. Irreversible trapping was more significant in HBQ Coreweld than in API X70 steel due to high amount of non-metallic inclusions in the weld microstructure.

Effective hydrogen diffusion coefficient was slightly higher for HBQ Coreweld than for API X70 steel. This was presumably due to the small grained microstructure of the former which, when aligned parallel to the diffusion direction, creates high diffusivity paths. The sub-surface hydrogen concentration was slightly higher for API X70 steel than for HBQ Coreweld.

The density of reversible trap sites for API X70 and HBQ Coreweld was 2.7 ・ 1019 sites cm−3 and 5.1 ・ 1019 sites cm−3 respectively. This is attributed to a higher density of inclusions, dislocations and grain boundaries in the weld metal than in the base metal. These lattice imperfections constitute reversible trap sites giving a higher value of traps’ density.

The binding energy of reversible trap sites was determined as 35.6 kJ mol−1 for API X70 steel and 32.7 kJ mol−1 for HBQ Coreweld. The nature of reversible traps in API X70 and HBQ weld was similar and the small difference was likely due to experimental scatter.