what effects temper martensite embrittlement have on fatigue crack nucleation and propagation?
Dr. Hou:
Temper embrittlement in high-strength martensitic steels is associated with the (co)segregation of alloying elements (such as Ni and Mn) and impurity elements (P and Si) to prior austenite grain boundaries during tempering or after welding (typically at around 500C); as such, it promotes decohesion in these boundaries, brittle intergranular fracture and a significant loss in fracture toughness. I examined the effect of this embrittlement on the fatigue-crack propagation behavior of the ultrahigh strength NiCr 300M steel (Si-modified 4340) many years ago (the paper is referenced below). I found that prior temper embrittlement substantially lowered resistance to fatigue-crack propagation, particularly at lower stress intensities approaching the fatigue threshold stress-intensity range for crack growth - this is the region that (more than often) dominates the life of a component - although (as is often the case), higher growth rate behavior was less affected. What is important to note though is that if hydrogen gas or water vapor is also present, temper embrittlement can act in a synergistic manner with any environmental activity, as all these mechanisms often focus their "attack" on these prior austenite boundaries.
The paper, if you're interested is Ritchie RO, Influence of impurity segregation on temper embrittlement and on slow fatigue crack growth and threshold behavior in 300-M high strength steel, Metall. Trans. A, 8A (7), 1131-40. July 1977.
ROR
Dr. Hou:
Temper embrittlement in high-strength martensitic steels is associated with the (co)segregation of alloying elements (such as Ni and Mn) and impurity elements (P and Si) to prior austenite grain boundaries during tempering or after welding (typically at around 500C); as such, it promotes decohesion in these boundaries, brittle intergranular fracture and a significant loss in fracture toughness. I examined the effect of this embrittlement on the fatigue-crack propagation behavior of the ultrahigh strength NiCr 300M steel (Si-modified 4340) many years ago (the paper is referenced below). I found that prior temper embrittlement substantially lowered resistance to fatigue-crack propagation, particularly at lower stress intensities approaching the fatigue threshold stress-intensity range for crack growth - this is the region that (more than often) dominates the life of a component - although (as is often the case), higher growth rate behavior was less affected. What is important to note though is that if hydrogen gas or water vapor is also present, temper embrittlement can act in a synergistic manner with any environmental activity, as all these mechanisms often focus their "attack" on these prior austenite boundaries.
The paper, if you're interested is Ritchie RO, Influence of impurity segregation on temper embrittlement and on slow fatigue crack growth and threshold behavior in 300-M high strength steel, Metall. Trans. A, 8A (7), 1131-40. July 1977.
ROR
It is a widespread issue with lots of variables. In low carbon steels (c
Dr. Hou:
I think that I answered you previously about the effects of temper embrittlement on fatigue-crack propagation in steels. However, with respect to your question about static modes such as intergranular cracking during fatigue crack growth at high growth rates (as the Kmax during the fatigue cycle approaches the overload fracture condition at, for example, KIc), overload fracture occurs when the entire fracture takes place by unstable (in this instance) intergranular cracking, whereas prior to that during the latter stages of fatigue, isolated grains fail intergranularly. Eventually when enough grains fail in this manner, overload fracture occurs.
This same scenario will take place with other static modes of fracture, such as transgranular cleavage cracks and microvoid coalescence. Depending upon the material, the occurrence of these static modes of fracture during fatigue cracking can also accelerate growth rates at high Kmax levels just prior to the onset of unstable fracture (overload failure).
You should note though that whereas these additional (monotonic) fracture mechanisms of fatigue are interesting academically, they play little role in determining the overall fatigue life of a component as the crack is growing too fast at this stage. The fatigue life is invariably dominated by crack growth at much lower growth rates, which is why near-threshold fatigue-crack growth is a much more important topic to study.
ROR
https://www.researchgate.net/publication/301348952_Licensed_Under_Creative_Commons_Attribution_CC_BY_Investigation_on_Mode_-_I_Fracture_Parameters_Using_Steel_Fibers_in_High_Strength_Concrete.
Article Licensed Under Creative Commons Attribution CC BY Investigat...
Data Investigation on Mode – I Fracture Parameters using steel fi...
Dr. Professor Ritchie:
fatigue crack initiated, extend to short crack length ,then long crack until failure.Does fatigue initiated crack have definite length? How it extend from initiated crack to short crack which then extend according to Paris power law?
what's physical meaning of fatigue endurance limit?does it mean no crack initiation?
Dr. Hou:
These are conceptually difficult questions, but here's my take on the problem:
1. Crack initiation is a perplexing concept because of the question that you just asked - how long must a crack be to be considered a crack? how long must it be to be initiated? It is almost a question of semantics! Engineers often describe crack initiation as when one can see a crack, but the resolution of the human eye is hardly the basis for a scientific criterion. In fracture mechanics methodologies, cracks/defects are presumed always to pre-exist, and so life calculations are based solely on how long these pre-existing cracks are and how long they will take to propagate until they result in overload fracture. I fully subscribe to this notion - it's clean, concise and does not rely on imponderables! It means that the major issue is not whether there is a crack there or not - it is simply how long this crack is!
However, fracture mechanics has its problems too. For example, linear elastic fracture mechanics (LEFM) is based on the Principle of Similitude which postulates that regardless of how long the crack is, if the stress intensity at the crack tip is the same, then it will behave the same, e.g., propagate at a specific rate. When cracks become short or small, this notion can break down and in many cases cause small cracks to grow faster than the equivalent large crack subjected to the same nominal stress intensity; moreover, they can often propagate at stress intensities below the measured (large-crack) fatigue thresholds. What is a small crack then? I have defined them as being (i) small compared to the scale of the microstructure (this is a continuum limitation); (ii) small compared to the extent of the plastic zone (this is a LEFM limitation), and (iii) small compared to the extent of crack closure (more generally, crack-tip shielding) in the wake of the crack tip. This is a complex problem though beyond the scope of this brief answer. However, if you're interested, I have written a few articles that might be of useful to read; these are S Suresh, RO Ritchie, The Propagation of Short Fatigue Cracks, International Metals Reviews, vol. 29 (6), Dec. 1984, pp. 445-476, and RO Ritchie, J Lankford, Small Fatigue Cracks: A Statement of the Problem and Potential Solutions, Materials Science and Engineering, vol. 84, 1986, pp. 11-16.
2. Your second question on what is the meaning of the "fatigue endurance limit" is also a hard one; the answer is almost certainly dependent on the nature of the material in question. In general though, I think that in most cases, it does not necessarily represent the alternating stress to initiate a crack. I believe that naturally initiated cracks or pre-existing ones can quickly start to propagate and then get locally arrested by local microstructural features, such as grain boundaries or particles. The fatigue limit represents the limiting stress for the crack to overcome these obstacles and allow the crack to propagate to failure.
ROR
Dr. Professor Ritchie:
There are many theories about fatigue carck initiation,and in high cycle fatigue , fatigue carck initiation occupy most of total life.notch and inclusion have much influence on fatigue life, especially on fatigue carck initiation. Do you think what’s the influence of notch and inclusion on fatigue?
cracks/defects are presumed always to pre-exist.Do they extend according to the principle of short crack in the early life?
Dr. Hou:
Notches are important in engineering structures because they are a source of stress (and/or strain) concentrations that can facilitate or exacerbate the nucleation of fatigue cracks. Similarly, inclusions are prime sites for fatigue crack initiation, often because they can be initially cracked as many inclusions are quite brittle (this is one of the reasons that fracture mechanics methodologies assume the presence of pre-existing cracks). This means that both notches and the presence of inclusions can have a large effect on the fatigue life of a component.
When such an incipient crack starts to propagate, this is essentially what we term "small crack growth" and this stage invariably dominates the overall life of the component. However, for fracture mechanics to be appropriate for the description of this initial crack growth, i.e., the growth rates can be characterized in terms of a global parameter such as the stress-intensity range, these cracks should be large enough so that their length is no longer comparable with the scale of the microstructural features, crack-tip plasticity and extent of crack-tip shielding. This is the paradox - small crack growth is the process that dominates the overall life of many structures yet it is the most complex to characterize in terms of standard fracture mechanics strategies.
ROR
Dr. Hou:
This is an excellent question and actually a difficult one to answer (as there are so many factors involved). In principal, the fatigue-crack growth threshold is the stress-intensity range below which crack growth occurs at vanishingly-small growth rates (essentially below which cracks do not grow) in a cracked sample, whereas the fatigue endurance strength is the alternating stress below which fatigue failure will not occur (as in the case of a fatigue limit) or will induce a life in excess of a certain value (such as >108 cycles in the case of the 108-cycle endurance strength), invariably in an unnotched sample. In simple terms, the first represents the limiting condition for fatigue crack growth, and the second the limiting condition for fatigue crack initiation (although the latter statement is not totally correct as even in unnotched specimens there are pre-existing cracks and the endurance strength can often represent the stress to get these pre-existing cracks growing by overcoming local microstructural arrest features such as boundaries).
The really interesting question though is how they are related, and the paradoxical answer to this is that often they are not, e.g., coarsening the grain size of a metallic alloys can lower the endurance strength yet at the same time increase the fatigue threshold. The reasons for these differences are many-fold and complex, but the simple answer is that the fatigue threshold represents a parameter pertinent to large cracks, i.e., large compared to the length-scales of microstructure and (generally) local plasticity, and is thus affected by (what I term) extrinsic mechanisms such as crack-tip shielding (mainly crack closure in fatigue), specifically at low load ratios. The endurance strength, conversely, is associated with smaller cracks, which are far less affected by crack closure phenomena, with their behavior controlled by intrinsic effects (for want of a better expression, by the inherent fatigue resistance of the material). As intrinsic and extrinsic effects rarely go "hand in hand", this is what can lead to the seemingly paradoxical differences in how material parameters such as yield strength and grain size lead to opposite changes in the fatigue threshold and fatigue endurance strength.
This is a messy subject because, like all issues pertaining to fatigue, there are so many factors which can play a role. I have tried to explain this in various of my publications; if you're interested, the attached publication may be one place to start.
ROR
Article Mechanisms of fatigue crack propagation in metals, ceramics ...
Dr. Hou:
You are essentially correct here. The fatigue endurance strength is more related to a small crack threshold, although the former is invariably measured as the alternating stress to cause failure or endure a specific number of cycles whereas the latter is measured as a stress intensity range above which a crack will begin to propagate. Both by the nature of their respective tests involve small cracks and are thus far less affected by phenomenon such as crack-tip shielding, in particular crack closure. So if you wish to seek any correlation between the endurance strength and a fatigue crack growth threshold stress intensity range, the small crack threshold would probably provide the closest fit.
ROR
Dr. Hou:
Clearly strength and ductility will influence the initiation and growth of fatigue cracks in steels - this is without question. For example, the strength will influence the total life, high-cycle fatigue S/N (Wohler) curve, as the fatigue limit, or endurance strength, is typically of the order of half the tensile strength. Similarly, for low-cycle fatigue, as described by the Coffin-Manson relationship, the ductility is effectively the constant in the Coffin-Manson law (essentially representing the extrapolation of the law to the value of the alternating plastic strain at 1/4 cycle). With respect to crack-propagation rates, the role of ductility is less definitive, although the strength is inversely related to the alternating crack-tip opening displacement (CTOD) which many believe is the most important parameter controlling the increment of crack advance per cycle in fatigue. Additionally, in steels, the higher the strength, the lower the extent of crack closure; for this reason the value of the fatigue crack propagation threshold is inversely proportional to the strength, especially at low load ratios.
However, the point to remember is that we should really be considering the cyclic strength and ductility, i.e., the value of the strength and the value of the ductility measured in a fatigue-cycled material (typically at roughly half life), although this is rarely done. The reason for this is the dislocation configurations set up under cyclic loading may be quite different to that established under monotonic loading. It is for this reason that some steels cyclically harden (e.g., as-quenched high-strength steels), some cyclically soften (e.g., quenched and tempered high-strength steels), and some initially soften and then harden (e.g., low-strength mild steels). However, since the cyclic strength and ductility properties are often not known and the effect of cyclic hardening/softening is generally not that large, i.e., typically less than roughly 20%, most correlations remain between cyclic fatigue properties and the monotonic strength and ductility.
Some of these issues for steels are discussed in one of my early papers, which is attached.
ROR
Article Influence of Microstructure on Near-Threshold Fatigue Crack ...
Dr. Hou:
Fatigue properties and strength and ductility do not necessarily correlate because they involve entirely different deformation and fracture mechanisms, as I mentioned in my previous response. However, if your are referring to the total fatigue life under high-cycle fatigue conditions, the S/N curve and the fatigue limit will nominally scale directly with strength, i.e., higher strength will in general confer a high fatigue limit and longer lives at a specific alternating stress. However, where this notion can become uncertain is that higher strength steels are more sensitive to stress concentrations, and this can also lead to premature failures. And so if you look at any general plot of fatigue limit as a function of strength (as for example in the ASM Metals Handbook), the general trend is for the fatigue limit to increase directly with strength but the degree of scatter becomes very much larger for high strength materials. This is clearly on of those "different situations" as you call them.
The bottom line is that higher strength should enhance the high-cycle fatigue life and higher ductility should enhance the low-cycle fatigue life, but these are only correlations as cyclic fatigue and tensile stress-strain (under monotonic loading) involve quite different deformation and fracture mechanisms.
ROR
Dr. Professor Ritchie:
The general theory of fatigue initiation , the persist slip band cause intrusion and extrusion and stress concentration at intrusion and extrusion,then initiate Fatigue crack . Is the intrusion and extrusion fatigue crack ? Or intrusion and extrusion just act as stress concentration, but how fatigue crack initiate at stress concentration?
Dr. Hou:
When is a crack a crack? A good question but it is one of semantics rather than engineering. Extrusions and intrusions can lead to fatigue crack initiation at a perfectly smooth surface. Somewhat like shuffling a pack of cards, alternating slip can lead to inward intrusions that can be the origin of fatigue cracks. They have been imaged - they are real, they can result in crack initiation from a smooth surface but to my mind they are not that an important source of fatigue crack initiation in real materials as the process of crack initiation is invariably associated with pre-existing defects or more generally imperfections. (I always regard them as being analogous to dislocation nucleation via Frank-Reed sources - another elegant mechanism which has been observed - but let's face it, most dislocations are nucleated at interfaces, at ledges in grain boundaries, etc.). I regard intrusions in a similar vein - most fatigue cracks initiate at local stress concentrations from imperfections at surfaces, from brittle inclusions in steels and many other engineering alloys, generally at or near the surface which crack or debond, from hard alpha grains that are locally aligned in titanium alloys, from etch pits in the presence of corrosive environments, or from local rubbing/corrosion at touching surfaces via fretting - these are the primary mechanisms of fatigue crack initiation in real materials, as opposed to some smooth-surfaced (often times electropolished) test specimen in the lab.
ROR
Dr. Hou:
The prime factor involved in the initiation of fatigue cracks is plasticity, specifically cyclic plasticity. The presence of a stress or strain concentration simply promotes such plasticity leading to the initiation of fatigue cracks. What is interesting here is that when the early fatigue failures first were recognized, for example in steel railway axles in the mid 1800s in Europe, the components failed at applied stresses less than the yield strength which really puzzled the design engineers as elastic failures would never be expected. What was happening, of course, was that stress concentrations were locally elevating the stresses above the yield strength resulting in local plasticity and hence the initiation of fatigue cracks, which then grew progressively until total failure ensued - this is exactly how all high-cycle fatigue failures occur.
ROR
Rob - Greetings. First, I enjoyed reading your discussion to the questions posed. Admire your patience. I was reading your paper on 300-M steel. You know that I do not subscribe to the plasticity induced crack closure. I just wrote a review paper on the fatigue crack growth resistance in the short crack growth regime to be published in MSEA journal. Coming back to your 300-M steel and the role of temper embrittlement - I like your last conclusion, for obvious reasons. '
'8) The effect of load ratio on crack propagation behavior is consistent with the environmental influence of hydrogen at low growth rates, and apparently inconsistent with crack closure concepts.'.
Regards - Sada
Hi Sada:
Great to hear from you.
I wrote that paper over 40 years ago, long before I appreciated the role of crack closure. I now think, as you know, that closure is a very important phenomenon during fatigue crack growth. But I agree with you that it's not necessarily plasticity-induced crack closure. Whether the latter mechanism exists or not is something that we can debate later, but for fatigue crack growth at near-threshold levels, that I was always interested in, plasticity-induced closure is not a relevant mechanism controlling behavior. In this near-threshold region, I have always believed that wedging phenomena, such as oxide-induced and especially roughness-induced crack closure, are the pertinent crack-tip shielding processes. This notion is consistent with many of the observed behaviors in this region, such as the effects of load ratio, coarse grain sizes, strength levels, etc., which can markedly influence the value of the fatigue threshold.
So I may have been correct in that last conclusion of mine from 1977, or it may simply reflect the "sins of my youth" (where the term "sins" refers to youthful inexperience!).
Cheers Rob
Omid - Please study Prof. Ritchie's paper. He has provided the answer in terms of the role of the temper embrittlement and also the role of H-embrittlement. Best Wishes.
For metallic aircraft components and structures the primary fatigue nucleation sites are stress concentrations, including manufacturing discontinuities. See Barter, Molent, Wanhill, International Journal of Fatigue (2012), 41, 11-22.
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering