Band 33, Heft 6 - Special issue: Fundamentals of environmentally induced cracking / Issue editors: A. K. Vasudevan, N. J. Henry Holroyd, Neville R. Moody and Ronald M. Latanision

Airframe structure is the core capability for all aviation operations, whether fixed or rotary wing, manned or unmanned, or ship-based or shore-based. Airframe materials are the underlying enabling technology for all air vehicle structures. Airframe maintainability is the primary availability and readiness driver for U.S. Naval Aviation. Traditionally, airframe structures are designed for immediate mechanical performance and loads-only structural response, and the degradation of properties over the life cycle and sustainment during operations are often an afterthought. Galvanic management and corrosion-resistant materials selection have never been done systematically as part of the structural analysis and lifing process. Additionally, the lack of true failure mechanism understanding of load path effects, crack initiation and growth behaviors, and nonuniform material response has often resulted in underdesign/overdesign, limits on new material insertion, costly life extension programs, and unexpected early in-service failures. Advances in structural and materials science for airframes offer significant opportunity for improvements in availability, readiness, reduced sustainment requirements, fatigue life enhancement, reduced weight and improved range, and enhanced design tools and standard practices. These science and technology advances would be realized in large part through the engineering and operational communities by facilitating mission profile-specific life prediction and operational requirement-tailored functionality, increasing operational readiness, reducing life-cycle costs, reducing logistics footprint, and lowering the maintenance burden on uniformed personnel. Durability can therefore be incorporated into the design and construction phase, where the largest return can be realized.

In this paper, I will attempt to summarize my experience of approximately 35 years on environmental fatigue damage, which I gained from industries outside Milan, Italy, and as a visiting scientist at the Naval Research Laboratory in Washington, DC. I will present a brief summary of my understanding of the topic of environmentally assisted fatigue of metals. About halfway in my career, I moved from the nuclear industry towards the oil and gas industry. This shift allowed me to understand that environmentally assisted cracking, which is an important topic for pipeline steels, is also similar to the nuclear pressure vessel steel community and that the problems are similar. At the same time, I became aware that the scientific communities are not working well to communicate the issues and challenges. People working in one industrial sector do not seem to appreciate the contribution from another industrial sector. This is understandable, as there are industrial rules and regulations. Because of such constraints, a knowledge gap is introduced into the technical community. This lack of cooperation slows down the progress of relevant research, especially nowadays, when we are facing increasingly complex service situations that may lead to failures. While trying to cope with such difficulties, I became interested in knowledge management. I learned that there is a need to build bridges between contemporary science, engineering, and business. We are facing big challenges in analyzing failures, and maybe it is time to do things differently and consolidate different disciplines of engineering with applications that can give us more reliable predictions. Organizations and companies worldwide are currently facing (a) increasing quantity of available information on material usage, (b) increasing number of people with high education in science and engineering, (c) increasing complexity of the failure problems to be solved, and (d) increasing lack of communication between institutions. It is estimated that the material loss in the US due to corrosion alone is approximately $300 billion per year. The necessity of knowing, understanding, and processing such complex data and problems is a very large task. Thus, there is a need for exploiting the technical information and using communication technology tools to help support teamwork, networking, using available experience and expertise, and, hopefully, international cooperation. Knowledge management involves a cultural change to stimulate a better use of all types of resources. Progress will be slow and difficult if we do not learn how to cope with this new situation.

The mechanisms of transgranular corrosion fatigue in age-hardened Al-Zn-Mg (-Cu) alloys are reviewed, with an emphasis on accounting for fractographic observations. The effects of cycle frequency, solution composition, temperature, and electrode potential on crack-growth rates (and striation spacing and appearance) at intermediate to high Δ K are discussed in particular. It is concluded that corrosion fatigue, resulting in cleavage-like {100} <110> cracking with extensive slip on {111} planes intersecting crack fronts, can best be explained by an adsorption-induced dislocation emission mechanism (involving weakening of interatomic bonds at crack tips by adsorbed hydrogen). Solute hydrogen ahead of cracks appears to play little, if any, role in facilitating cleavage-like cracking, and the mechanisms based on decohesion are at odds with the locally high strains around cracks and the formation of nanovoids ahead of cracks.

There are complex issues remaining to be resolved before environment-assisted cracking models can be included in structural mechanics programs that are currently used to analyze mechanical fatigue crack growth. Considered here is the effect of cyclic frequency on fracture mode transitions that occur during corrosion fatigue of high-strength Al-Zn-Mg-Cu alloys. These alloys are used in civilian and military aircraft applications where they are exposed to detrimental aqueous saline environments. A previously developed “critical hydrogen at a critical distance” crack growth model is used to rationalize the observed transitions in crack-path fracture-modes, from intergranular (IG), to brittle transgranular (BTG), to ductile transgranular (DTG), as the alternating stress intensity factor and cyclic frequency increase beyond critical values. Corrosion fatigue crack growth rate data obtained on Al-Zn-Mg-Cu alloy 7017-T651 base metal and heat-affected “white zone” metal tested in aqueous NaCl solutions over a frequency range of 0.01–70 Hz are analyzed. For the white zone metal, dependence of the critical crack velocity on the critical frequency, at which the IG-BTG transition occurs, undergoes an abrupt reversal as the critical frequency increases above about 0.1 Hz. Mechanisms potentially responsible for this change in frequency dependency are discussed in the context of the critical hydrogen model. The transition from low to intermediate frequency behavior is speculated to be due to a change in the critical distance from microstructural control, for frequencies at or below 0.1 Hz, to control by the critical hydrogen criterion at higher frequencies. The low frequency behavior is discussed relative to the transition from static load stress corrosion cracking to low frequency corrosion fatigue, which occurs as cyclic frequency increases above zero.

Crack initiation and growth behavior in 4340 steel under static and cyclic loads in various environments are examined using the two-parameter approach developed by the authors. It is shown that environmental effects can be time dependent, stress dependent, or both, depending on the amplitude, peak stress, and aggressiveness of the environment. It is shown that the tests under vacuum are required to establish the reference state for the quantification of the chemical driving forces involved in crack initiation and growth in aggressive environments. Failure diagrams are proposed by extending the Kitagawa-Takahashi diagram for fatigue. It is shown that these can be used to quantify the mechanical and chemical internal stresses required to cause both initiation and growth of an incipient crack in a smooth as well as notched specimens. Conditions for propagating and nonpropagating incipient cracks that are initiated under stress concentrations are also discussed, taking the example of the 4340 steel data available.

It is well recognized that environment has a significant role on the failure of mechanically loaded structures. In most cases of cyclic loading, fatigue crack growth (FCG) behavior exhibits lower threshold and faster growth rate in air than in vacuum. It is well documented that the effect of loading frequency on FCG behavior can be more pronounced in aggressive environment/material systems. This is seen in the K max term of the FCG. On the other hand, a weak dependence of FCG behavior with R ratio in inert environment indicates that a crack extension is governed mainly by Δ K . Existing experimental data indicate that the actual crack extension per cycle is associated with the rising part of the load cycle than the unloading part. In this paper, the synergetic role of environment and mechanical loading on crack growth behavior is considered to see their roles on FCG. In this article, we attempt to model how crack extension interplays between a crack-tip opening and crack-tip blunting angle associated with the applied load and environment, respectively. To support such a model for discussion, we have selected limited FCG data taken from literature corresponding to different environments ranging from vacuum to air and NaCl solution for a number of alloys and with different specimens geometries. We are also not discussing innate mechanisms for each alloy, due to space concerns.

When Al-Cu-Mg (AA2024-T351) alloy plate is immersed in 0.53 m NaCl solution, intergranular corrosion occurs, which initially takes the form of narrow fissures at the surface. Although growth of these fissures is initially rapid, it soon slows and the penetration depth is limited. However, these intergranular fissures can give rise to intergranular subcritical crack growth and eventual catastrophic fracture if specimens are strained slowly in laboratory air after pre-exposure to the NaCl solution. In our experiments, tensile specimens of the alloy were strained at strain rates in the range 10 -7 –10 -4 /s, and the cracking propensity was found to increase with pre-exposure times up to around 300 h. However, cracking was much reduced with longer pre-exposure times and became negligible after 500 h. This rather unexpected result can be explained by the observation that the intergranular corrosion fissure becomes increasingly broad and un-crack-like with extended pre-exposure times, failing to meet the prerequisite initiation conditions required for intergranular stress corrosion cracking. A further unexpected result was that straining specimens in the same 0.53 m NaCl solution after pre-exposure produced no intergranular crack growth. To investigate this further, tensile tests were conducted under electrochemical control, and this showed that the free corrosion potential in the test solution was too low in 0.53 m NaCl to allow cracking to proceed.

This work provides a broad overview of the systematic experimental studies conducted in our group to understand the factors governing the transition to a more occluded corrosion front during the repassivation of Al alloys, manifested by the appearance of an inflection during anodic polarization into the active region in NaCl solutions. The collected data and the derived empirical relationships enabled us to discern the thermodynamic and kinetic aspects controlling the transition to more occluded local damage. From the thermodynamic standpoint, the potential at the inflection, namely, the pit transition potential E ptp , can be considered as the thermodynamic driving force of Al dissolution in the acidified pit-like solution in contact with a freshly created surface. The associated current density i ptp and the steepness of the potential decrease with current below E ptp are determined by interfacial electrochemical kinetics. More importantly, all these properties are influenced by electrochemical and metallurgical events, including residual stresses. These findings point to the electrochemical approach in combination with external mechanical load as a promising tool for investigating environmentally assisted cracking, in particular, crack nucleation and non-steady crack tip processes.

Transmission electron microscopy (TEM) was employed to investigate the dissolution behavior of nanocrystalline grain boundary T 1 precipitates in Al-3Cu-2Li. These grain boundary T 1 plates exhibit an orientation relation with matrix, with the (1-11)α-Al parallel to (0001)T1 and [022]α-Al parallel to [10-10]T1, which is similar to the orientation relationship of T 1 plates formed within grains. TEM studies showed that these grain boundary T 1 plates react readily in moist air. As a result of the localized dissolution, the Cu-rich clusters form onto T 1 , which is consistent with the localized dissolution behavior observed in nanocrystalline S phase in Al-Cu-Mg.

Stress corrosion characterization is analyzed in two alloys, 5083 and binary AL-3Li, both having anodic grain boundary precipitates. The 5083 commercial alloy forms AL 3 Mg 2 (β phase), whereas the AL-3Li alloys have ALLi (δ phase). These are intermetallic phases that form on the grain boundaries and grow in size and area fraction with heat treatment time. When exposed to the NaCl solution, they anodically dissolve and form grain boundary cracks under an applied stress. A brief electrochemical description of these phases reacting with NaCl solution is described. The crack initiation occurs when the appropriate local environmental conditions at the crack tip and stress state is established. Overall, crack initiation mechanism seems to be a combination of mostly anodic dissolution coupled with hydrogen-assisted cracking.

Notch sensitivity effects under environmentally assisted cracking (EAC) conditions have been recently quantified considering the tolerance to short cracks that may start at their tips and become nonpropagating after growing for a while, a behavior that depends on the stress gradients ahead of the notch tips and on the basic material resistances to crack initiation and propagation inside an aggressive medium. This model can provide a powerful alternative design tool for the pass/nonpass criterion traditionally used to deal with such mechanical-chemical problems, since it properly considers and quantifies the stress analysis issues that affect them. The model uses the analogy between the notch sensitivity behavior under fatigue and under EAC conditions, so it quantifies how the stress gradient around the notch tips affects the tolerance to mechanically short cracks that depart from there, considering the characteristics of the loading and of the notch geometry, as well as the basic material properties inside the given environment, expressed by its EAC resistances to crack initiation from a smooth surface S EAC and to crack propagation K IEAC , without the need for any data fitting parameter. Moreover, since this model has been validated by proper tests, it can be used to propose a defect-tolerant design criterion under EAC conditions that includes the unavoidable notch effects always present in actual structural components.

The vulnerability of tempered martensitic steel to hydrogen embrittlement (HE) has attracted attention from a number of researchers. Although utilizing carbide precipitation is one of effective methods to improve HE resistance, few studies have focused on the effects of carbide characteristics, such as the chemical composition and morphology of carbide. This work clarifies the role of Mo carbide and V carbide in the HE behavior of tempered martensitic steels with four steels whose chemical composition was carefully controlled. The beneficial effect of carbides is discussed in terms of hydrogen trapping and fracture mode. The low amount of trapped hydrogen and undissolved carbide led to excellent HE resistance of Mo carbides compared to V carbides. In addition, the superior mechanical performance of Cr-Mo steel was also interpreted by the effect of Cr addition as well as Mo carbides.

Stress corrosion cracks of all types are characterised by extensive crack branching, and this is frequently used as the key failure analysis characteristic to identify this type of cracking. For aluminium alloys, stress corrosion cracking (SCC) is almost exclusively an intergranular failure mechanism. For plate and extruded components, this had led to the development of test procedures using double cantilever beam and compact tension precracked specimens that rely on the pancake grain shape to constrain cracking, so that fracture mechanics can be applied to the analysis of stress intensity and crack velocity and the evolution of a characteristic performance curve. We have used X-ray computed tomography to examine in detail SCC in aluminium alloys in three dimensions for the first time. We have found that crack branching limits the stress intensity at the crack tip as the applied stress is shared amongst a number of cracks that are held together by uncracked ligaments. We propose that the plateau region observed in the v - K curve is an artefact due to crack branching, and at the crack tips of the many crack branches, cracking essentially occurs at constant K almost irrespective of the crack length. We have amplified the crack branching effect by examining a sample where the long axis of the pancake grains was inclined to the applied stressing direction. Our results have profound implications for the future use of precracked specimens for SCC susceptibility testing and the interpretation of results from these tests.

The critical energy release rate, defined as G c =2 γ + γ p (where γ is the cohesive energy and γ p is the plastic work), is widely used as a macroscopic fracture criterion. During embrittlement of aluminum by gallium, impurity segregation is localized to a small region along the grain boundary and the bulk plastic properties are unaffected. Yet, the critical energy release rate (which is predominantly described by the plastic work [ γ p ] since γ p >> γ ) significantly decreases. In this work, we recognize that as the cohesive energy ( γ ) decreases during corrosion due to an increase in impurity concentration, the stress needed at the notch tip to form the crack decreases, and this, in turn, decreases the plastic work by reducing dislocation emission at the notch tip. We study two different models proposed in the past that can capture this dependence of γ p on γ during liquid metal (Ga) embrittlement of aluminum alloy (Al 7075). The parameters in these models are computed from first principles atomistic calculations and recent experiments. We compare and contrast these models on their ability to describe various aspects of embrittlement such as fracture toughness, K IC , and subcritical value of stress intensity, K Iscc . Extension of the approach to predict threshold fatigue crack initiation in Al7075 is suggested.

Material strengthening and embrittlement are controlled by interactions between dislocations and hydrogen that alter the observed deformation mechanisms. In this work, we used an energetics approach to differentiate two fundamental stress corrosion mechanisms in iron, namely, hydrogen-enhanced localized plasticity and hydrogen-enhanced decohesion. Considering the small-scale yielding condition, we use a discrete dislocation framework with line dislocations to simulate the crack-tip plastic behavior. The crack growth was modeled using the change in surface energies (cohesive zone laws) due to hydrogen segregation. The changes in the surface energies as a function of hydrogen concentration are computed using atomistic simulations. Results indicate that, when hydrogen concentrations are low, crack growth occurs by alternating mechanisms of cleavage and slip. However, as the hydrogen concentrations increased above some critical value, the crack grows predominately by the cleavage-based decohesion process.

The adoption of Ti-6Al-4V titanium alloy is widespread in advanced engineering fields, such as the aerospace, automotive, and maritime sectors, due to its remarkable mechanical characteristics and proven corrosion resistance. However, literature studies and field failures have underlined that the Ti-6Al-4V is susceptible to the stress corrosion cracking (SCC) and the corrosion fatigue phenomena in methanol-water solutions. Although several studies focused on the SCC behavior of this particular titanium alloy, recent corrosion fatigue experiments ( R =0.1) on Ti-6Al-4V specimens in a water-methanol environment have highlighted a sensible decrease in maximum applied stress of up to 50% and a high sensitivity even to low methanol concentrations. In the present work, an experimental quasi-static SCC test campaign in water-methanol mixture at different concentrations has been conducted on flat dogbone Ti-6Al-4V specimens. The results have been compared with recent corrosion fatigue experiments to characterize the different damage mechanisms at play and decouple mechanical and chemical driving forces. The micromechanical behavior of the material has been analyzed, highlighting the role of the microstructure in the SCC mechanism. Failure surfaces have been investigated with scanning electron and optical microscopy.

The role of the excessive metal vacancy generation in transgranular stress corrosion cracking (TGSCC) of pure copper was studied in relation to crack initiation and growth mechanisms. Electrochemically polarized specimens were strained at room temperature in 0.3 m NaNO 2 solution. Cation vacancy redistribution under applied anodic polarization took place at the p-n junction of the duplex Cu 2 O oxide film, resulting in cation vacancy transport in the p-type layer of the oxide to the metal substrate interface and in excessive metal vacancy accumulation in the metal substrate. The rapid initiation of TGSCC at room temperature in the cold-worked oxygen-free high-conductivity copper samples occurred during slow strain rate tests under anodic polarization in 0.3 m NaNO 2 solution. The duplex Cu 2 O oxide film structure formed by cathodic deposition, before applying the anodic polarization, was characterized with X-ray diffraction and field emission scanning electron microscopy/electron backscatter diffraction/energy-dispersive spectroscopy. The redistribution of cation vacancies in the p-type oxide phase due to the anodic polarization and the metal vacancy accumulation in the copper substrate interface was studied by using positron annihilation spectroscopy.

Fracture propagation involves the coupling of many length scales ranging from the sample loading geometry to the molecular level. In brittle materials, the length scales of the damage process zone are reduced to a submicrometric scale and the coupling with the macroscopic scale is expected to be the domain of linear elastic fracture mechanics. However, although 2D elastic analyses are generally adequate to describe the sample deformation at macroscopic scales, local investigations of failure mechanisms at the sample-free surface require the use of 3D mechanical tools due to the crack front local curvature and to the corner point singularities at the intersection between the crack front and the external surfaces of the sample. We present here a thorough multiscale investigation of the slow crack growth of a sharp crack in oxide glasses in the stress-corrosion regime, combining experimental and numerical analyses of the displacement fields from the millimeter scale to the nanoscale range. The principal aim of the study is identifying the length and time scales of the mechanisms of damage and interaction between water and glass, which have been the subject of an extensive debate in the last decades.

Constant load tests of high-strength carbon steels with different micro-alloying using strengths in the range of 1000–1400 MPa were performed at ambient temperature under continuous electrochemical hydrogen charging. Hydrogen markedly affects delayed fracture of all the studied steels. Fractography of the studied steels shows that fracture mechanism depends on the chemical composition of the studied steels and hydrogen-induced cracking exhibits intergranular or transgranular character occurring often in the form of hydrogen flakes. The size and chemical composition of non-metallic inclusions are analyzed by scanning electron microscopy and energy-dispersive X-ray spectroscopy. Hydrogen-induced cracking initiates at TiN/TiC particles in steels with Ti alloying. Crack paths are studied with electron backscatter diffraction mapping to analyze crack initiation and growth. The thermal desorption spectroscopy method is used to analyze the distribution of hydrogen in the trapping sites. The mechanisms of hydrogen effects on fracture of high-strength steels are discussed.

This article deals with the risk of environmentally assisted cracking of steel structures that are kept under cathodic protection (CP). The experimental results collected on both hydrogen diffusion and hydrogen embrittlement (HE) of high-strength low-alloy (HSLA) steels under CP are discussed. Hydrogen diffusion was evaluated by permeation experiments and a scanning photoelectrochemical current technique, as a function of microstructure orientation, on both loaded and unloaded specimens. HE tests were carried out under constant load, slow strain rate (SSR tests), and slow bending conditions. Tests were also carried out on several grades of HSLA steel having different microstructures. The results confirm that HE in artificial seawater under CP can occur in steel with tensile yield strength in the range of 400–660 MPa only in the presence of high cathodic polarization and continuous plastically straining conditions. HE susceptibility increased with increasing applied cathodic polarization and with decreasing strain rate. HE susceptibility of the rolled steels at relatively high strain rate (10 -4 to 10 -5 s -1 ) increased with the hydrogen diffusion coefficient. Similar results were observed in terms of the HE contribution to corrosion fatigue crack growth rate. High-temperature-tempered martensitic steels showed a lower HE susceptibility.

A multiscale analysis has been conducted on hydrogen-induced intergranular cracking at ambient temperature in medium strength (840 MPa) Ni-Cr steel with antimony, tin, and phosphorous segregation. Combining first-principles calculations and fracture mechanics experiments, a multiscale relationship between threshold stress intensity factor ( K th ) and cohesive energy of grain boundary (the ideal work of interfacial separation, 2 γ int ) was revealed. The K th was found to decrease rapidly under a certain threshold of 2 γ int , where the 2 γ int decreases mainly by mobile hydrogen segregation on fracture surfaces. This segregation is considered to arise during formation of the fracture surfaces under thermodynamic equilibrium in slow crack growth. The resulting strong decohesion probably makes it difficult to emit dislocations at the microcrack tip region, leading to a large reduction in stress intensity factor. Our analysis based on this mobile hydrogen decohesion demonstrates that the K th decreases dramatically within a low and narrow range of hydrogen content in iron lattice in high-strength steels.