Extrusions and intrusions in fatigued metals. Part 1. State of the art and history†

†In 2003, it was 100 years since J.A. Ewing and J.C.W. Humfrey first documented the nature of early surface fatigue damage in Swedish iron in their paper The Fracture of Metals under Repeated Alternations of Stress published in the journal Philosophical Transactions of the Royal Society of London, Series A [1]. Let this work (Part 1 and 2) commemorate their pioneering effort to understand the physical mechanisms of early fatigue damage.

Abstract

Current state and historical progress in experimental and theoretical studies of surface relief appertaining to persistent slip bands (PSBs) and leading to fatigue crack initiation in cyclically deformed metals is presented as a thorough critical overview. A comprehensive inventory of microscopic techniques used for this study is tabulated chronologically with emphasis to their applicability to polycrystals. The most relevant experimental characteristics concerning surface relief evolution, namely the form of extrusions and intrusions in single- and polycrystalline materials, are surveyed. Theoretical models and computational simulations of extrusion and intrusion formation and fatigue crack initiation are critically reviewed.

†In 2003, it was 100 years since J.A. Ewing and J.C.W. Humfrey first documented the nature of early surface fatigue damage in Swedish iron in their paper The Fracture of Metals under Repeated Alternations of Stress published in the journal Philosophical Transactions of the Royal Society of London, Series A [1]. Let this work (Part 1 and 2) commemorate their pioneering effort to understand the physical mechanisms of early fatigue damage.

Keywords:

• persistent slip band
• persistent slip marking
• extrusion
• intrusion
• fatigue crack initiation

Acknowledgements

The authors are grateful to Professor L.M. Brown (his group's model) and to Dr. E. Hieckmann (formerly Thiele) (internal stresses within PSBs) for useful discussions. They are also indebted to an anonymous referee for his critical and stimulating comments on the manuscript. Support of the present work by the grants No. 106/06/1096 and No. 101/07/1500 from the Grant Agency of the Czech Republic and by the research project No. AV0Z 20410507 and the grant No. 1QS200410502 from the Academy of Sciences of the Czech Republic is acknowledged.

Notes

†In 2003, it was 100 years since J.A. Ewing and J.C.W. Humfrey first documented the nature of early surface fatigue damage in Swedish iron in their paper The Fracture of Metals under Repeated Alternations of Stress published in the journal Philosophical Transactions of the Royal Society of London, Series A [1]. Let this work (Part 1 and 2) commemorate their pioneering effort to understand the physical mechanisms of early fatigue damage.

Notes

1. The term “intrusion” for a fine crevice-like depression within the PSM has been firstly used by Hull 3. However, some contemporary authors do not mention extrusion–intrusion pairs and instead they refer to microcrack initiation in the wake of extrusion (see Section 4). Although discussions of fatigue crack initiation still have difficulty with the definition of a nucleated crack, some experimental results clearly suggest that there is a difference between intrusion and crack nucleus (see Section 6).

2. In addition to the typical extrusion–intrusion pairs observed most frequently, isolated intrusions were also detected 38,43,73.

3. Only Ma and Laird 87 reported different types of protrusions in copper single crystals. After a large number of cycles, they observed that some macro-PSBs, typically 50 µm in thickness, are associated with net inward displacement (i.e. negative protrusions and/or encroachments) on the surface of the single crystal. Negative protrusions were also observed on the front-surface in the rectangular aluminium single crystals 58 and in cylindrical copper single crystals at the position where the primary Burgers vector represents a tangent to the specimen surface 89,90.

4. Very little is known about the macroscopic stress state within PSB lamellae. The only two experimental studies on macroscopical internal stresses within PSBs are conflicting. Ogin and Brown 25 made birefringence observations in optically transparent ionic AgCl single crystals subjected to cyclic loading, and their results indicated that PSBs and matrix are in opposite macroscopic stress states, which were reversed in sign in each half-cycle. Buque et al. 105 monitored Barkhausen noise in nickel single crystals cyclically deformed into the saturation stage and they concluded that permanent compressive internal stresses with small cyclic changes exist in PSBs, which do not reverse sign during the loading half-cycle.

5. It is worthwhile mentioning in this context that a similar, unusual configuration of voids and fatigue cracks growing perpendicularly to the PSBs was also observed by Atkinson (see 103, pp.194–195) in copper single crystal, dispersion-hardened with SiO₂ particles and having the same crystal geometry as shown in Figure 4a. According to Brown 103, these cracks are probably caused by the accumulation of vacancy dipoles at the incoherent interface of the hardening particles (compare also with the model by Hsiung and Stoloff 114, Section 4.2.3).

6. At this point, it should be pointed out that the formation of intrusions by analogous mechanism sketched in Figure 5 is excluded in the EGM model 17,18.

7. Lin and Ito 115 intuitively suggested that the initial stresses in P and Q may be provided by a row of interstitial dipoles along slice R, in the same visual fashion as considered by EGM (see Figure 5a), i.e. slice R is in a state of compression. The only experimental evidence documenting indirectly the particular distribution of initial shear stresses in P, Q and R is that by Wood and Bender 119 who studied slip band displacement across artificial scratches in copper subjected to alternate torsion (see also 99,117).

8. The second extreme case of the most favorably oriented grain considered by Tanaka and Mura 125 and later by Chan 128 is that with the slip direction parallel to the specimen surface. In this case, however, no extrusions or intrusions are formed by dislocation accumulation during cycling and, thus, it is not presented here.

9. It should be noted that copper single crystals used for the study at the mentioned temperature interval were pre-fatigued with a plastic strain amplitude of ±2 × 10⁻³ or ±1 × 10⁻³ at room temperature to about half the saturation stress 44.

10. A similar ‘ratcheting’ mechanism whereby extrusion–intrusion pairs are formed at pre-existing stage I cracks has been already considered in a more general form by Neumann et al. 132.

11. The grain size considered in the computer simulations by Brinckmann and Van der Giessen 140,141 is very close to the minimum grain size of ∼1 µm or less, which was found to be the limit for formation of PSBs with a ladder structure 145 and/or any dislocation patterning 146.

12. Only transcrystalline crack initiation mode is considered in this section. In addition to this mode, PSBs or PSMs can also play a role in crack initiation along grain or twin boundaries, see e.g. 38,76,147,148. For detail description of different mechanisms proposed for intercrystalline crack initiation mode, see e.g. 12,14,15.

13. Although there is no consensus on the definition of a just nucleated crack so far, detailed observation of surface relief topography of macro-PSMs in copper single crystals using section micromilling technique by Hunsche and Neumann 42,75 suggests that intrusions and crack nuclei are distinctively different in two aspects: (1) their geometrical shapes are different (vertex angle 30° or 0°, respectively) and (2) intrusion formation does not depend on the environment, whereas crack formation and growth do strongly depend on the environment.

14. Although Brown and co-workers in their model do not distinguish specifically between individual PSBs and macro-PSBs, their model can also be applied to macro-PSBs, reflecting their specific dislocation structure observed in the early and later stages of fatigue life (see above) 103.