Thirty-five years in environmentally assisted cracking in Italy: a point of view

1 Early observations

In 1946, after the Second World War, a research center called the Centro lnformationi Studi Esperienze (CISE) was created in Milan, Italy, to support research in nuclear power generation. There I was involved starting in 1978 in research activities focused in the field of environmentally assisted cracking (EAC), more particularly in the environmental fatigue testing of reactor pressure vessel (RPV) steels. At the beginning of my training, I spent some time in 1979 at the Naval Research Laboratory (NRL) in Washington, DC, where corrosion fatigue was studied on steels and titanium and aluminum alloys using fracture mechanics principles developed by George Irwin. I worked with several people at NRL, and when I returned to Milan, a brand-new laboratory was built at the CISE to work on EAC. I had the opportunity to take part in the international scientific community working on EAC and to complete some test programs in the subject. Unfortunately, in 1987, the Italian Government stopped building nuclear reactors, and the decision was made also to stop any research activity relating to the use of nuclear energy. Such a decision ended the existence of the CISE in 1996.

Meanwhile, in 1994, a few of us moved to the research laboratories of the Italian Oil and Gas Company. The subject of the research I was involved in was mainly stress corrosion cracking (SCC) of pipeline steel, internal corrosion of pipelines, and coupling of corrosion with flow dynamics. This is also where I became involved in a project on knowledge management. In 2004, after 10 years, due to the reorganization of research in the Company, I moved to exploration and production and stopped to do direct research work in the laboratory; however, I am still involved in EAC and I try to keep contact with the research community.

2 EAC

EAC is defined as a damage form where acute defects (cracks) nucleate and grow due to the combined action of aggressive environment, applied load, and the metallurgy of the material. EAC can be the result of different damage mechanisms, for instance, hydrogen diffusion in the metal and/or anodic dissolution coupled with stress.

However, when corrosion occurs, anodic and cathodic processes are present at the same time, and it is sometimes difficult to distinguish between different damage mechanisms.

In Figure 1, the main parameters involved in EAC are graphically represented. In EAC, the applied load may be both static and dynamic, as schematically shown in Figure 2, so that the damage form can be called environmental fatigue, SCC, or hydrogen embrittlement. The damage mechanisms are often dependent on strain rate, ε dot, either applied externally or at the crack tip.

Figure 1:

EAC parameters.

Figure 2:

Strain rate effect.

In the early days, we were involved in the study of the environmental fatigue of RPV steel and focused on different types of tests, all on the same material and similar environments of demineralized water with or without some oxygen at two temperature levels.

Using large CT specimens (2 inches wide), we measured crack tip chemistry by inserting a potential probe and a pH electrode inside the crack (Gabetta, 1982; Gabetta & Rizzi, 1983) for measurement. A summary of test results is shown in Figure 3. Both oxygen depletion and crack tip acidification were observed (Edwards & Schuitemaker, 1987; MacDonald & Urquidi-MacDonald, 1988).

Figure 3:

Potential and pH measurements at the crack tip: test method and results.

The effect of loading (fatigue) was included in the study using the superposition model (Austen, 1983; Gabetta, 1987) at different loading frequency (Vosikovsky, 1975) and is shown schematically in Figure 4.

Figure 4:

Schematic of the superposition model for corrosion fatigue.

This model takes into account two different effects of the environment on fatigue. If general corrosion is superimposed onto fatigue loading, the Paris law line [plot of cyclic crack growth rate (da/dN) vs. the stress intensity factor range (ΔK)] is moved towards higher crack growth velocities. This is true corrosion fatigue. If damage due to SCC is superimposed on true corrosion fatigue, a component with constant growth rate can be added to the Paris law to describe crack growth (Gabetta & Torronen, 1986). In this case, different frequency values are associated with different plateau crack growth rate values, as in the following equation:

( d a / d N ) s c f = ( d a / d N ) t c f + ( d a / d t ) s c c × 1 / f r e q u e n c y .

This model agrees with striation spacing measurements (Gabetta, Rinaldi, & Lucia, 1992). Ductile and/or brittle striations are associated with the true corrosion fatigue and stress corrosion fatigue components of the crack growth rate, respectively. Striation measurements do confirm the existence of both mechanisms in different sections of the fracture surface, often close to each other. A method to measure striation spacing was developed by Claudia Rinaldi at the CISE (Gabetta, Pozzi, & Rinaldi, 1990). Examples are shown in Figures 5 and 6.

Figure 5:

Example of ductile striations and spacing measurements.

Figure 6:

Examples of brittle striation spacing compared with crack growth rate at different frequency.

The value of spacing varies as a function of loading frequency for brittle striations due to stress corrosion mechanisms; ductile striation spacing is a function of the applied stress intensity factor range and falls on the Paris law for true corrosion fatigue.

Transgranular SCC can be observed in pipelines transporting hydrocarbons, following the well-known spectrum proposed by Parkins (1963) and summarized in Figure 7.

Figure 7:

Stress corrosion spectrum in pipelines.

In a buried pipeline, an increasing applied load can be generated by a moving landslide, as summarized in Figure 8. The SCC component is associated with a plateau crack growth rate value.

Figure 8:

Superposition model for transgranular stress corrosion in pipelines.

Transgranular SCC in pipeline steel can be easily compared with the brittle component of corrosion fatigue. We applied the superposition model to pipeline steel and succeeded in reproducing, in a laboratory sample, the fracture surface observed in the field. However, oil and gas researchers prefer to use the term “near-neutral pH SCC.” In my opinion, this description can cause ambiguity. On the contrary, in that community, transgranular SCC is mainly associated with hydrogen embrittlement, whereas anodic dissolution is associated with intergranular damage and an active path. The two cracking mechanisms are observed in the field, under different conditions, and can be easily differentiated. The discussion on cracking mechanisms is more open in the group of people working in the environmental fatigue of RPV steel (Cullen, Gabetta, & Hänninen, 1985; Shoji, Komai, Abe, & Nakajima, 1986). My participation in the International Cyclic Crack Growth Rate Cooperative Group was a very stimulating experience, and the discussion on the subject (anodic dissolution, hydrogen embrittlement, slip and dissolution, and so on; see Figure 9) is still active and open to debate.

Figure 9:

SCC proposed mechanisms.

3 Comments on acronyms and definitions

Damage mechanisms observed in carbon steel in contact with aqueous solutions are usually related to hydrogen-induced cracking and anodic dissolution. However, it seems that names, definitions, and opinions on data interpretation are widely different and depend not only on the observation of phenomena but also on the bias of the scientist who is studying such a problem. Nowadays, the amount of information and the analytical means available to researchers is very large, yet concrete experiments are lacking to clearly separate the damage mechanisms.

In the following, one can see a list of definitions-acronyms about hydrogen embrittlement, stress corrosion, and cracking mechanisms. This list is only partial but contains acronyms not present in the NACE Style Manual (NACE International, 2008) or in the NACE/ASTM G193 Standard (NACE/ASTM G193-12d, 2013).

One may suggest that work is needed to distinguish between types of damage and damage mechanisms. Moreover, there is a need for simplification to avoid having too many different and overlapping definitions, which has diluted the damage analysis. As our knowledge increases and with all of the new analytical information that has become available, we develop a deeper understanding of damage mechanisms. However, in going deeper and trying to identify those mechanisms, we tend to lose the site of the damage that occurs in practice and to forget that the goal should be how to avoid damage in service components.

As a concluding remark, I would like to suggest assembling a task force to list all definitions and mechanisms and damage formulas, with the objective of obtaining consensus and promoting the use of a lower number of acronyms with a physical meaning that can be accepted by many. It can be a large task. People with experience in understanding corrosion failures, component design, inspection, etc., should be gathered to address this problem.