Characterizing the Fatigue Behavior of the

High-Entropy ‘Cantor’ Alloy CrCoMnFeNi

by

Keli Vieno Siukola Thurston

A dissertation submitted in partial satisfaction

of the requirements for the degree of
Doctor of Philosophy

in

Engineering – Materials Science and Engineering

in the

GRADUATE DIVISIONS
of the
UNIVERSITY OF CALIFORIA, BERKELEY

Committee in charge:
Professor Robert O. Ritchie, Chair
Professor Andrew Minor
Professor Lisa Pruitt

Fall 2018
 Abstract

Characterizing the Fatigue Behavior of the

High-Entropy “Cantor” Alloy CrCoMnFeNi

by

Keli Vieno Siukola Thurston

Doctor of Philosophy in Engineering – Materials Science and Engineering

University of California, Berkeley

Professor Robert O. Ritchie, Chair

High-entropy alloys (HEAs) have spurred great interested in the in the materials
community in recent years, primarily due to the intriguing properties certain members
have been shown to display and due to the novel approach to alloy design they represent.
Of this family of multi-component alloy systems, the near-equiatomic five component
“Cantor” alloy CrCoMnFeNi is of particular note due to the remarkable mechanical
properties this alloy has been found to display which are only enhanced as the
temperature is lowered into the cryogenic regime. Despite the interest in this alloy
system, little work has to date been conducted that characterize the cyclic fatigue loading
behavior of this alloy or of its compositional variants. Here I examine the damage-tolerant
fatigue behavior of the Cantor alloy as well as the effect temperature and load ratio have
in changing this behavior and the underlying mechanisms that may be responsible for
the observed shifts. These testing conditions encompassed three temperature regimes:
293 K, 198 K, and 77 K; additionally, the impact of increased load ratio R was surveyed
within each temperature regime. Fatigue testing in the threshold and linear portion of the
Paris regime revealed a temperature dependence of the fatigue behavior of the Cantor
alloy; as temperature was decreased into the cryogenic regime, the fatigue curve was
found to shift toward higher ΔK, indicating a higher resistance to fatigue-crack
propagation at lower temperature. Additionally, it was observed that higher load ratios
impacted this resistance negatively, causing a shift towards lower ΔK with increasing R-
ratio. Post-testing, a series of mechanistic studies were conducted to investigate the
underlying cause for this observed shift. Analyses of crack closure measurements, crack-
path morphology, and fractography provided strong evidence for roughness-induced
crack closure as the dominant mechanism at play.

1
To those who are full of determination

(especially you, MLST)

i
 Acknowledgements

This journey has been long and arduous but not one I have had to travel alone. Rather,
many have contributed to make this project a successful one, more than I can detail fully.
Whether through direct collaboration or indirect support, you have my deepest gratitude.

However, I would like spend these few words to recognize and thank some of you in
greater detail. Chiefly, Professor Ritchie and my fellow group members. Professor Ritchie
and Bernd, you two taught me so much and acted as wonderful mentors both within the
laboratory space and without. My Ritchie group PhD cohort, you’re the best. Whether it
was giving me a break so I could go take a nap during a long dry-ice test, helping prep
for the exams, or providing constructive feedback, I could always count on you. Perhaps
we’ll collaborate in the future; regardless, I’ve made several lasting friendships here.

Additionally, I would like the thank the many undergraduate and master’s students
who worked directly with me to get these tests up and running. Edwin, Jessica, and
Julian, thank you so much for the work you put in; it really means a lot to me. Mingxi
and Andrew, we may not have worked directly as much, but I’ve loved having you as
colleagues as well, sharing ideas, stories, cookies.

Next, moving away from the laboratory space, I must acknowledge my other
graduate school colleagues and friends. While our work may differ widely (from other
MSE subfields to environmental biology to economics/finance to pottery), our successes
and stumbles along the path to the degree are similar. You fellow travelers were
wonderful companions, cheering me during the high points, supporting me during the
low, and ensuring I took some time every so often to enjoy the vistas along the way.

Kiitos paljon, my family; I can’t express how important you all are to me in a paltry
couple of sentences. Leila and Brian, my mom and dad, you shaped me to be the person
I am today, particularly evident in my love of books, crafting, and science. Matti, you’ve
always being there for me; I couldn’t ask for a better brother. Grandma Bev and Grandpa
Phil, you’ve always supported me wholeheartedly in all my endeavors and continue to
inspire me to this day. You all mean the world to me, and I’m so happy to have you all in
my life!

Sincerest thanks to all my colleagues, collaborators, friends, and family. This

dissertation would not exist without you. And finally, I’d like to acknowledge my
funding sources. Much of my dissertation is based upon work supported by the National
Science Foundation Graduate Research Fellowship and by the Mechanical Behavior of
Materials Program (KC13) at Lawrence Berkeley National Laboratory.
ii
 Table of Contents

Abstract............................................................................................................................................ 1

Acknowledgements ...................................................................................................................... ii

List of Figures ................................................................................................................................ iv

List of Tables .................................................................................................................................. viii

Chapter 1: Introduction ................................................................................................................ 1

Chapter 2: Background ................................................................................................................. 3

2.1 Growing Interest in High-Entropy Alloys .............................................................. 2
2.2 Fatigue Studies of HEAs ............................................................................................ 6
2.3 Mechanistic Studies .................................................................................................... 11
2.4 Discussion .................................................................................................................... 12

Chapter 3: Fatigue Characterization of the Cantor Alloy ..................................................... 17

3.1 Materials and Methods .............................................................................................. 18
3.2 Results .......................................................................................................................... 20

Chapter 4: Mechanistic Studies of Fatigue in the Cantor Alloy ......................................... 27

4.1 Threshold Studies ....................................................................................................... 27
4.2 Crack Closure Studies ................................................................................................ 28
4.3 Crack Path Studies ...................................................................................................... 30
4.4 Fractography ................................................................................................................ 34
4.5 Summary ...................................................................................................................... 37

Chapter 5: Discussion .................................................................................................................. 39

5.1 Comparison to other alloy systems .......................................................................... 39
5.2 Suggested future work ............................................................................................... 41

Chapter 6: Conclusions ................................................................................................................ 43

References ...................................................................................................................................... 44

iii
 List of Figures
2.1 Overview of recently examined HEAs. The grouping into the above
categories is based on the compositional similarity of the included
HEAs......................................................................................................................... 4

2.2 Stress range over cycles to failure for three separately processed
Al0.5CoCrFeNi alloy systems at room temperature............................................ 8

2.3 Fatigue crack growth rate over the stress intensity range for AlCrFeNi2 Cu
and Al0.2CrFeNiTi0.2 by Seifi et al. Tests were performed with 20 Hz at room
temperature with R-ratios of 0.1, 0.3, and 0.7 ..........................................…...... 10

2.4 Plot depicting the relationship between the fatigue ratios and the ultimate
tensile stresses of Al0.5CoCrCuFeNi and other common engineering
alloys......................................................................................................................... 14

2.5 Comparison of the fatigue endurance limits at 107 cycles between

Al0.5CoCrCuFeNi and other conventional alloys. The HEA was found to
outperform most of the common engineering materials though also
displayed wide scatter depending on the precise processing conditions...... 14

3.1 Images of sample preparation and mounting in the Instron load frame.
Clockwise from left: Wheatstone bridge consisting of three 350 Ω for signal
amplification, sample with attached strain gauge and full wiring, side view
of prepared samples, and sample mounted in load frame............................... 19

3.2 Fatigue-crack growth behavior of the CrMnFeCoNi alloy, tested at a load

ratio R of 0.1 at 293 K and 198 K. A direct comparison of the fatigue-crack
growth curves, as a function of the stress-intensity range, ΔK, at both
temperatures tested on disc-shaped compact-tension (DC(T)) samples at a
frequency of f = 25 Hz reveal a clear shift in the fatigue threshold, increasing
some 30%, from ~4.8 MPa√m to ~6.3 MPa√m, as the temperature was
reduced from ambient to cryogenic conditions.................................................. 21

3.3 High magnification SEM of fracture surfaces at comparable ΔK ~ 20

MPa√m. (a),(b) depict 198 K tested sample in the region ΔK ~ 22.7 MPa√m
while (c),(d) depict the ambient 293 K tested sample in the region ΔK ~ 21.6
MPa√m to allow comparison between the two sets of images under similar
loading conditions.................................................................................................. 23

iv
3.4 Fatigue crack growth rates as a function of the stress-intensity range for the
Cantor alloy under various testing conditions. Testing was conducted over
three temperature ranges between ambient (298 K) and liquid-nitrogen
temperature (77 K) at R = 0.1, 0.4, and 0.7 at 25 Hz. Tests at 77 K were limited
to the upper region of the mid-growth rate regime while testing at 293 K
and 198 K continued until the fatigue threshold stress intensity was
reached. A direct comparison of the crack-propagation curves
characterizing these various testing parameters indicates a clear upwards
shift in the value of the ΔKTH threshold as both temperature and load ratio
were decreased........................................................................................................ 24

3.5 Example of fatigue striations observed in the high growth rate region of
samples tested at 198 K. (a) and (b) depict samples fatigued at R = 0.4 and
0.7 respectively. Each image depicts striations of width on the order of the
expected fatigue crack growth rate. .................................................................... 26

4.1 Plot of threshold ΔKTH and KMAX values of samples tested at 293 K and 198
K with respect to R-ratio........................................................................................ 28

4.2 Crack path characterization of the CrMnFeCoNi alloy fatigue tested at 293
K and R = 0.1. Back-scattered electron (BSE) images and electron back-
scatter diffraction (EBSD) scans taken at the mid-thickness of the samples
near the threshold at ΔK = 4.8 MPa√m, specifically from the wake of the
crack at ΔK values of ~6 and ~7 MPa√m, clearly show transgranular fracture
as the dominant crack propagation mode. Images taken at the crack tip
indicate that the crack propagated through the grains and recrystallization
twins and not along their boundaries. EBSD scans taken from the wake of
the propagating crack show plastically deformed region along the crack
flanks resulting from physical contact between the mating crack surfaces
and the lower strength of the alloy at room temperature................................. 31

4.3 Crack path characterization of the CrMnFeCoNi alloy fatigue tested at 198
K and R = 0.1. In contrast to the behavior at room temperature (Fig. 4.2),
both BSE images and EBSD scans revealed intergranular fracture as the
dominant crack propagation mechanism at 198 K. The boundaries of the
recrystallization twins do not appear to influence the path of the crack tip,
as can be seen near the threshold at ΔK = 6.3 MPa√m and at higher growth
rates at ΔK values of ~7.5 MPa√m and ~8.5 MPa√m; this is believed to be a
result of the higher strength of the material at 198 K, such that plastically
deformed regions along the propagated crack do not to occur....................... 32

v
4.4 EBSD imaging of the crack path in the high stress intensity regions of LN2
tested samples. To examine the potential role of nano-twinning in the shift
of the fatigue curve in cryogenically-tested CrMnFeCoNi, detailed EBSD
scans were taken at the crack-tip region, the area at which the largest stress
intensities were experienced. Images (a) and (b) correspond to imaging of
samples subjected to R = 0.1 with ΔK ≈ 29 MPa√m and R = 0.7 with ΔK ≈ 20
MPa√m respectively with crack propagation direction indicated by inset
arrow. Image (c) shows a higher resolution portion of the crack path of the
R = 0.7 sample as indicated by the inset highlight box...................................... 33

4.5 Fractographic analysis of CrMnFeCoNi samples tested at R = 0.1 in the 293

K and 198 K temperature regimes. Fracture surface characterization of
samples tested at 293 K show mainly intergranular crack propagation (a-c)
with some minor transgranular failure regions (inset of (c)). The highly
serrated fracture surface features covering most of the intergranular failure
region (insets of (b) and (c) are likely associated with cyclic slip steps
resulting from dislocation motion by planar slip [10]. Similarly, the near-
threshold region at 198 K (d) shows planar slip as major deformation
mechanism at this temperature whereas at higher growth rates (e,f) crack
propagation occurs mainly through intergranular fracture. The white arrow
in (d) and the insets of (e) indicate polyhedral features that likely show the
ends of deformation twins formed perpendicular to the grain boundaries
into the grains neighboring the crack front......................................................... 35

4.6 Fractographic analysis of CrMnFeCoNi samples tested at R = 0.4 at 198 K at

two regions subject to different stress intensity ranges. (a) depicts a region
subject to moderately high ΔK ~ 9 MPa√m while (b) shows a region subject
to high ΔK ~ 17 MPa√m.......................................................................................... 36

4.7 Overview of fractography over all nine testing conditions at a moderately

high ΔK ~ 12 MPa√m. Images of fracture surfaces at 77 K (left column) show
evidence of post-testing oxidation though the general underlying shapes
correlate well with the trends observed between 198 K and 293 K tested
samples. Higher R seemed to correlate with increased surface plasticity and
sharpness of features, as predicted. Additionally, a greater degree of flatter,
facet-like features with lower temperature continued to be observed at this
higher ΔK though not to the same degree at in Fig. 4.5, as expected by the
observed increasing degree of transgranular fracture behavior with higher
ΔK.............................................................................................................................. 37

vi
5.1 Comparison of the fatigue-crack growth behavior of the CrMnFeCoNi high-
entropy alloy with austenitic stainless steels and TWIP steels (at 293 K
unless otherwise noted). The Cantor alloy displays fatigue behavior most
similar to twinning-induced plasticity (TWIP) steels [61-65]. Austenitic
stainless steels with comparable microstructure show slightly lower
threshold with similar Paris slopes of approximately 3.5-4.5. The Al-
containing HEAs demonstrated both higher Paris-regime slopes and
fatigue thresholds [35], which may be attributed largely to the as-cast
microstructure in combination with the small sample sizes that were used
during testing.......................................................................................................... 40

vii
 List of Tables
2.1 Overview of the total-life data regarding recently examined HEAs . ............. 7

2.2 Test results produced by Seifi et al. for the AlCrFeNi2Cu and the
Al0.2CrFeNiTi0.2 alloy for different R-ratios.......................................................... 10

3.1 Summary of fatigue crack growth and threshold data, as shown in Fig.
3.4.............................................................................................................................. 24

4.1 Summary of fatigue crack growth data with accompanying crack closure
data............................................................................................................................ 29

5.1 Comparison of fatigue crack growth testing parameters and results for data
shown in Fig. 5.1..................................................................................................... 40

viii
 Chapter 1
Introduction
Our society’s continual progress and growth drives a persistent demand for novel
materials to fulfill roles and applications previously only dreamed of, step by step
removing the ‘fiction’ from science. Metallic alloys continue to play a fundamental role
in the basis of many of these technologies. As such, advancements in metallurgy and alloy
design remain key steps on this path of scientific and engineering advancement. One of
the most exciting subfields of metallurgical to emerge over the last decade or so has been
the study of the “high entropy” alloy (HEA) family. This subfield has enjoyed enormous
growth over the last few years, with each year seeing hundreds of more papers being
published ranging from new members to the HEA family to characterizing their
properties to proving their feasibility in practical applications.

This interest is well-founded as HEAs represent a major shift in approach to alloy

design, one of humanities oldest technologies. Where conventional alloys consist of one
or two majority elements, HEAs consist of five or more, often in near-equiatomic
compositions, leading to a far greater number of potential alloy compositions and, in
turn, vast potential for hitherto unseen material properties and behaviors. Properties of
interest are as varied as the potential applications of metallic alloys, including but not
limited to electrical, thermal, and mechanical; in this work, I will be focusing on the
mechanical behavior of one particular HEA, specifically the fatigue behavior of the
Cantor alloy.

While the Cantor alloy CrMnFeCoNi is attributed as the first described HEA and
thus oldest member of this family of metals, it nevertheless remains one of the most
promising in terms of its mechanical and structural properties, exhibiting excellent
strength and toughness that only improves as the temperature is lowered into the
cryogenic regime, disobeying the typical strength-toughness trade-off displayed by most
metals at low temperature. However, as well-documented as many of this alloys
properties have been, little work has been done to characterize the behavior of this alloy
or any of its compositional variants under cyclic loading conditions, vital information to
know before this alloy may be put into practical use; here I seek to bridge this gap.

1
 Chapter 2
Background
High-entropy alloys (HEAs) represent a new approach to alloy design that has taken
the metallurgical world by storm over the past few years as new metallic alloys have been
found to exhibit unique combinations of properties; moreover, there is seemingly an
unlimited number of potential new compositions that have yet to be explored. One
particular HEA and its variants that have shown exceptional promise as structural
materials is the so-called Cantor alloy, which is a single-phase, face-centered cubic solid-
solution with an equiatomic CrMnFeCoNi composition. CrCoNi-based medium- and
high-entropy alloys have been shown to have unprecedented combinations of strength,
ductility and fracture toughness which are further improved at cryogenic temperatures;
however, few studies have investigated their fatigue properties, either in terms of total
stress/strain life or their crack propagation behavior. In general, these initial studies have
indicated a relative similarity of the fatigue properties of the Cantor HEA variants with
traditional steels of comparable strength and toughness. The precise energy dispersion
mechanisms at play, however, remain contested and appear to differ with composition;
current theories range from increased nano-scale twinning to crack closure effects
although more studies are required to further explore the role of these phenomena. While
some initial computer modeling studies have been conducted, the current lack of
experimental fatigue studies on HEAs suggests a pressing need to verify and improve
these models. More importantly, as fatigue is the most dominant failure mode for metallic
structural materials, there is a more general need to mechanistically understand the cyclic
deformation and fatigue failure properties of HEAs, both from a perspective of their
potential use in practical applications and to help guide future alloy design.

2.1. GROWING INTEREST IN HIGH-ENTROPY ALLOYS

In the continuing quest for new and superior structural materials, high-entropy alloys
(HEAs) present a fascinating proposition. First developed in the early 2000’s by Cantor et
al. [1] and Yeh et al. [2], HEAs represent an unbounded set of new metallic alloys,
encompassing, in principle, stable (or meta-stable) solid solutions of at least five major
constituent elements in nominal equiatomic concentrations. This is in stark contrast to
traditional metallic alloys which comprise one (or two) dominant elements, such as iron
in steels or nickel in superalloys, with small percentages of additional elements. The
unusually high number of constituent elements has suggested to many the possibility of
unusual mechanisms, such as local chemical ordering, sluggish diffusion, in situ phase
transformation, and so forth; indeed, the fact that most of these potential new alloys have

2
never been made, analyzed or even contemplated previously has raised the additional
possibility of finding materials with unprecedented combinations of mechanical and
functional properties [3]. As society and technology advances, the constituent materials
for these technologies must advance alongside them. This is especially true for structural
materials, which are used primarily for their mechanical properties such as strength,
toughness, and hardness. Structural materials are used everywhere: from the steel and
concrete that comprises much of our infrastructure, to the composites and aluminum that
our airframes are built from. Within this class of materials, metals have long dominated
due to an excellent combination of factors. They tend to have excellent mechanical
properties in terms of strength and toughness as well as exhibiting plasticity instead of
catastrophic failure that increases their factor of safety and reliability in applications
where extended use is needed. Furthermore, they are easily available with a wide range
of tunable properties because of their ability to alloy with other metals and heat treatment
cycles to maximize mechanical performance. Combining this all with their ability to be
formed and machined in a variety of complex shapes makes them a difficult candidate to
contend with for future materials.

The name “high-entropy alloy” originates from the early hypothesis that the high
configurational entropy of these multi-element solutions would overwhelm the enthalpy
of phase formation and thus be the key to permitting their stabilization into stable single-
phase materials [4,5]. However, more recent studies, such as those conducted by Otto et
al., indicate that the degree of configurational entropy exhibited by a solution is not the
sole criteria; mixtures displaying high configurational entropies do not always form
stable single-phase solid solutions [4]. Several studies have postulated that
microstructural stability is primarily affected by entropic stabilization effects [5]. As
such, the phrases used to describe this class of alloys have expanded to include ‘multi-
principal elements’ or ‘complex concentrated alloys’; however, the term ‘high-entropy
alloys’ still prevails as most enduring within the materials communities and is expected
to the one cemented into the metallurgical lexicon [5]. Figure 2.1 exhibits the breath of
solid solutions compositions that have come to be defined included in the category ‘high-
entropy alloys’ with more members continually being added as new compositions are
discovered and the definition broadens further; for example, some four-component alloy
systems have been described as HEAs while others, such as three-component systems,
have been classified as ‘medium-entropy alloys’ [7].

3
 Figure 2.1: Overview of recently examined HEAs. The grouping into the above categories is based on
the compositional similarity of the included HEAs [6].

While HEAs remained little more than a processing curiosity in the decade following
their discovery, these alloy systems have gained significantly more attention over the past
few years as initial investigations into the mechanical properties of the then most
prominent HEA, the equiatomic CrMnFeCoNi ‘Cantor alloy’, yielded exceptional results.
Early monotonic testing of this particular composition revealed high tensile strengths up
to 1.28 GPa combined with high ductility (~60-70%) and fracture toughness values over
200 MPa√m [1,4,8,9]. Even more notable, the mechanical behavior of the Cantor alloy was
found, at decreasing temperatures below ambient, not to degrade but rather to further
improve without exhibiting the typical strength-ductility tradeoff observed in most
metallic materials [10,11]. This trend continues well into the cryogenic temperature
regime, with the Cantor alloy demonstrating mechanical behavior comparable, and, in
some cases, exceeding the properties of state-of-the-art cryogenic steels [9,12].

Motivated by the exceptional mechanical properties of the Cantor alloy, there has
been a significant resurgence in research to characterize other promising and prominent
HEAs. Recent studies, for example, with an aluminum variant of the Cantor alloy,
AlCrFeCoNi, found it to display yield and fracture strengths of 1450 MPa and 2960 MPa,
respectively [13], values which with further modification with titanium-enrichment were

4
increased to 2.26 GPa and 3.14 GPa [14]. Other studies have led to new HEAs that
demonstrate high hardness, high compressive yield strength, and outstanding wear
resistance comparable to ferrous alloys with similar hardness that outperform SUJ2 wear-
resistant steels [2,15-21]. More exotic compositions, such as NbSiTaTiZr, have been found
to display thermal stability with excellent recrystallization behavior coupled with
excellent oxidation and corrosion resistance [9,22,23]. With fifty or so usable elements in
the periodic table, the number of new HEAs to be investigated, in terms of merely
equiatomic five-component systems, runs into several tens of millions, and so probing
this vast compositional space for new and useful alloys represents a major challenge,
involving theoretical techniques, e.g., first-principles DFT or CALPHAD procedures,
and/or machine learning algorithms, coupled with high-throughput experimental
techniques (although the best ways to perform this daunting task remains unclear).
However, despite the current feverish activities in the field of HEAs, one critical feature
in terms of their mechanical performance remains under investigated, and that is failure
under alternating loads - fatigue, despite its crucial role in the practical application of
these engineering alloys.

The role of fatigue in mechanical failure first began to be appreciated in the mid-1880s
with the failure of railroad axles; just under a hundred years later, the problem of fatigue
came to even greater attention with the premature failures of the De Havilland Comets
airliners, the first pressurized commercial jet aircraft, which resulted from fatigue cracks,
growing in the aluminum alloy fuselage, that had initiated from the corners of the square
windows. Now fatigue is recognized as a critical mode of fracture accounting for more
than 80% of all in service failure involving accidents ranging from the mundane to the
catastrophic [24,25]. Cyclic loading conditions can occur in many environments; as such,
a proper understanding of a material’s fatigue properties is vital in any engineering
application that may touch on human lives. Proper characterization of subcritical crack-
growth behavior can be used to construct accurate modern lifetime predictions of critical
parts under predicted cyclic loading conditions, thus allowing both calculation of safety
margins for a given material and a basis for comparison with other potential candidate
materials, aiding and improving future design [26]. Further, fatigue calculations can
determine the proper maintenance intervals for critical parts, allowing cost-saving longer
intervals between preventative maintenance while ensuring operator and user safety
[27,28].

Many of the catastrophic mechanical failures that have occurred can be traced to
misunderstood or undiagnosed fatigue failure due to cyclic loading conditions over long
intervals of subcritical loading. Yet despite the importance for understanding the fatigue
behavior for materials that may be used in a structural capacity, few HEAs have been

5
investigated in this capacity either in terms of fatigue lifetime or in crack propagation
behavior [29,30].

2.2. FATIGUE STUDIES OF HEAs

The studies of the fatigue properties of the HEA family as currently documented in
the literature remain quite sparse. However, these examinations that have been
conducted encompass a broad range of methods and techniques from total-life
predictions to crack growth propagation rate analysis to computational modeling.

2.2.1. Total-life approach studies

The first published work that focused specifically on fatigue testing of HEAs was
conducted by Hemphill et al. on a cast Al0.5CoCrCuFeNi alloy that displayed a two-phase
face-centered cubic (fcc) structure with a major α-fcc phase and a copper-rich β-fcc phase
[31]. The material was prepared by arc melting elements of ≥99% purity and annealing at
1000 oC for 6 h before being water quenched and cold rolled. Four-point bend specimens
of size 25 x 3 x 3 mm were then constructed and tested in ambient conditions under cyclic
loading at 10 Hz until failure or 10-7 cycles at an R-ratio, defined as the minimum divided
by the maximum stress of the cyclic load conditions, of 0.1. The maximum stress levels
varied between 540 – 1250 MPa with the upper boundary close to the examined yield
stress of 1284 MPa. The related Al7.5Cr22.5Fe35Mn20Ni15 alloy, a lower cost alternative, was
later processed and tested under equivalent conditions [31,32].

The test results revealed a wide scatter over the tested stress levels with an endurance
limit between 540 - 945 MPa for Al0.5CoCrCuFeNi and 540 - 630 MPa for the
Al7.5Cr22.5Fe35Mn20Ni15 alloy as listed in Table 2.1. The authors found no correlation
between the occurred scatter and the orientation and morphology of the phases. Upon
further investigation, they observed a trend related to the number of initial defects,
particularly aluminum oxide-rich particles in the material; specimens that contained a
lower number of defects in general performed better, with longer fatigue lifetimes [31].
The non-uniformity of defects within the entire sample set was hypothesized to be
attributable to the casting process, allowing the increased introduction of phase
segregation, inclusions, and microcracking in the material at sides of the casting which
may then become exacerbated during the cold-rolling processes. Indeed, examination of
the fracture surfaces showed that microcracks nucleated from the surface due to stress
peaks introduced by the roughness and nucleated perpendicular to the load direction
afterward. As the method of the weakest link suggests, the cracks initiated at the defects
on the surface. Besides the high dependency of the fatigue life on initial defects, a linear

6
correlation between the ultimate tensile stress and the endurance limit was found with
an approximate factor of 0.5; these results lend support for the theory that the high
fatigue-limits of HEAs can be attributed to their high ultimate stress values [31].

Table 2.2: Overview of the total-life data regarding recently examined HEAs
Material R-Ratio Microstructure Fatigue limit [Mpa] Fatigue ratio
0.1 as-cast 540-945 0.402-0.703
Al0.5CoCrCuFeNi [31,32]
0.1 as-cast (modified) 312-454 0.29
Al7.5Cr22.5Fe35Ni15Mn [32] 0.1 as-cast 540-630 -
Al0.1CoCuFeNi [34] -1 coarse-grained 200-210 0.36

The two-phase Al0.5CoCrCuFeNi was further investigated by Tang et al. [33]. To

observe and quantify the impact of oxide inclusions and microcracks on the high scatter
previously observed by Hemphill et al., three ingots were processed; two ingots were
produced using commercial-purity elements (I + II) while the third was formed of high-
purity elements (III). These ingots were repeatedly melted and solidified under vacuum
to promote homogenization of the material before annealing and quenching under
similar conditions previously conducted by Hemphill et al. Prior to the cold-rolling
process, shrinkage pores and macro-segregations were removed from the surfaces of
ingot II and III. Samples were machined, and testing conducted as previously described
by Hemphill et al. [31].

The endurance limit of the S-N curve was found to be at 383 ±71 MPa and a fatigue
limit to tensile strength ratio of 0.29; test results are shown in Table 2.1. Comparisons
between the sample groups manufactured from the three ingots revealed that scatter in
the results was significantly decreased through the use of high-purity elements and
removal of macro-defects prior to testing as shown in Figure 2.2.

7
Figure 2.2: Stress range over cycles to failure for three separately processed 𝐴𝑙 . 𝐶𝑜𝐶𝑟𝐹𝑒𝑁𝑖 alloy systems
at room temperature. The tests were performed with 10 Hz and an R-ratio of 0.1. Not failed specimens
were marked with an arrow, indicating their probably higher lifetime. The lines show the results of the
random endurance-limit fatigue model. Note the high scatter present in Condition 1 samples, similar to
that which Hemphill observed. The endurance limit of the Condition 3 samples was found to be ~383
MPa [34].

Two additional copperless compositional variants of aluminum-containing HEA

family have also recently been scrutinized for practical application in high-fatigue
environments. The solid mechanical properties exhibited by Al 0.1CoCrFeNi has caught
the attention of the medical community and examined as a potential material for stents
for cardiovascular applications [25]. The studies performed by Alagarsamy et al. differ
from those by Hemphill et al. and Tang et al. in two critical ways beyond the
compositional differences: here, the material was annealed at 1000 oC after the casting
and cold-rolling process to eliminate the cast microstructure, and the testing R-ratio was
-1 vs. 0.1. These differences make direct comparison difficult. Regardless, the coarse-
grained Al0.1CoCrFeNi was found to demonstrate a solid endurance limit of 200-210 MPa
with little scatter; in comparison, 316L stainless steel demonstrates similar mechanical

8
properties to CoCrFeNiMn, yet only exhibits an endurance limit of 184 MPa at a stress
ratio of -1 [34].

A more recent study conducted by Liu et al. of Al0.3CoCrFeNi furthered the discussion
of the relationship between the processing and fatigue properties of HEAs through
examination of the role of grain size, specifically of an ultrafine-grained microstructure,
and performance under fully reversible bending fatigue loading [35]. Implementing a
method similar to Alagarsamy et al., the cast and solutionized samples were subjected to
cold rolling to ~84% reduction and annealed at 700 oC for 36 hr to achieve a three-phase
material with the dominant FCC phase exhibiting an average grain size of 0.71 µm. As
with Alagarsamy et al., samples were subjected to cyclic tension testing at an R-ratio of -
1. Under these conditions, the fatigue runout of the HEA’s S-N curve was found to be 450
MPa with a fatigue runout to tensile strength ratio of 0.43, the best performance yet
reported for HEA systems tested in S-N fatigue [35]. As the composition remains similar,
this data suggests the ultrafine grained microstructure plays a significant role in its
superior performance.

In general, the total-life approach results generated by Hemphill et al. [31,32], Tang et
al. [33], Alagarsamy et al. [25], and Liu et al. [35] indicate the aluminum-containing Cantor
variants display favorable endurance limits comparable or superior to steels and other
conventional alloys under the appropriate processing conditions, with much of the
scatter and sensitivity in data collection linked to flaws potentially linked elemental
purity and to the casting and/or cold-rolling conditions.

2.2.2. Crack Propagation studies

Seeking to characterize the fatigue crack propagation in HEAs, Seifi et al. investigated
the impact of cyclic loading on the Al0.2CrFeNiTi0.2 and the AlCrFeNi2Cu alloy systems
[36]. Both as-cast alloys were found to be multiphasic, dominated by two phases
exhibiting face-centered cubic (fcc) and body-centered cubic (bcc) structures interwoven
in a dendritic/interdendritic manner. To generate these samples, high purity elements
were vacuum levitation melted and drop cast into a cylindrical mold before being
machined into 3 x 5 x 47 mm bend bar samples. Samples were then subjected to cyclic
testing at 20 Hz in ambient conditions with crack growth monitored using the direct
current potential drop method. Additionally, Seifi et al. investigated the impact of load
ratio during testing, performing tests at R-ratios = 0.1, 0.3, and 0.7. Results are
summarized in Table 2.2 and Figure 2.3.

9
Table 2.2: Test results produced by Seifi et al. for the 𝐴𝑙𝐶𝑟𝐹𝑒𝑁𝑖 𝐶𝑢 and the 𝐴𝑙 . 𝐶𝑟𝐹𝑒𝑁𝑖𝑇𝑖 . alloy for
different R-ratios. Tests were performed with 20 Hz at room temperature [34].

Figure 2.3: Fatigue crack growth rate over the stress intensity range for 𝐴𝑙𝐶𝑟𝐹𝑒𝑁𝑖 𝐶𝑢 and 𝐴𝑙 . 𝐶𝑟𝐹𝑒𝑁𝑖𝑇𝑖 .
by Seifi et al. Tests were performed with 20 Hz at room temperature with R-ratios of 0.1, 0.3, and 0.7. (a)
The fatigue threshold for 𝐴𝑙𝐶𝑟𝐹𝑒𝑁𝑖 𝐶𝑢 tested under these conditions was found to be ∆𝐾 ~ 17 𝑀𝑃𝑎√𝑚,
∆𝐾 ~ 5 𝑀𝑃𝑎√𝑚, and ∆𝐾 ~ 7 𝑀𝑃𝑎√𝑚, respectively. Those of 𝐴𝑙 . 𝐶𝑟𝐹𝑒𝑁𝑖𝑇𝑖 . were similar, as shown in
(b): ∆𝐾 ~ 16 𝑀𝑃𝑎√𝑚, ∆𝐾 ~ 7 𝑀𝑃𝑎√𝑚, and ∆𝐾 ~ 5 𝑀𝑃𝑎√𝑚 at the corresponding R-ratios [34].

Both alloy systems were found to exhibit remarkably high fatigue thresholds of 16
MPa√m and 17 MPa√m for Al0.2CrFeNiTi0.2 and AlCrFeNi2 Cu respectively at R-ratio = 0.1
with Paris slopes within the typical range of metals (3-6). These curves were both found
to shift lower with increasing R-ratio, leading to a decrease in fatigue thresholds to 7
MPa√m and 5 MPa√m respectively and a significant increase in the Paris slopes up to 25.8
for the more-brittle Al0.2CrFeNiTi0.2 [34]. These characteristics coupled with observation
of fracture surfaces that revealed a combination of ductile and brittle behavior lead the
authors to draw parallels between the behavior of these two alloy systems and other
metallic alloys, intermetallics, and metal matrix composite systems.

2.3. MECHANISTIC STUDIES

While the quantitative studies indicated fracture behaviors comparable to other

commonly used metallic alloys proving the suitability of HEAs in structural capacities,
further work was conducted to determine the mechanistic behavior responsible for the

10
observed fracture behavior. In addition to broadening our understanding of these alloys,
characterization of the fracture mechanisms within the HEAs may allow for more refined
and pointed generation of future alloy systems.

2.3.1. Nanotwinning

One most discussed mechanisms of interest in the field of structural HEAs is

nanotwinning and the contribution it may make in the remarkable deformation behavior
displayed by HEAs. Twinning is regularly observed in these materials in the form of post-
processing recrystallization twins, a favorable phenomenon due to the low stacking fault
energy of these multi-component systems. While deformation-induced nanotwinning
has been observed in certain strain-activated conditions, the role nanotwinning may play
is a topic of contention.

Transmission electron microscopy (TEM) imaging conducted by Tang et al. on post-

fatigue samples revealed dense tangles of dislocations and nanotwins within the material
[33]. Deformation nanotwinning would result in more twin boundaries, leading in turn
to additional work hardening as they act as high-angle grain boundaries and the crack
propagation was found to be crystallographic. The resulting decreased stress
concentration increases the strength of the alloy and thus the fatigue endurance of the
material. As such, nanotwinning was theorized to serve as the main deformation
mechanism leading to the high endurance limits of this dual phase HEA as the low
stacking fault energy due to Suzuki interactions as well as the high strength of HEAs
favorably affect the formation of nanotwins [33]. Niendorf et al. further corroborate this
theory of nanotwinning acting as an important deformation mechanism, particularly
under in cryogenic conditions [36].

The nanotwinning observed by Tang et al. was found to consist of both deformation-
induced nanotwins and processing-generated nanotwins; these observations align with
the results of Hemphill et al. who found a large number of nanotwins present in the
material after the cold rolling process [31-33]. Furthermore, Liu et al. attributed the large,
uniform elongation observed in the tensile behavior to the ultrafine grain Al 0.3CoCrFeNi
to the dynamic Hall-Petch effect resulting from nanotwinning behavior in the FCC matrix
phase of the alloy [34]. Liu et al. suggest this twinning propensity may further result in
the excellent fatigue properties observed in the material by reducing the crack growth
rate through nanotwin crack bridging in addition to the general work hardening-assisted
fatigue life enhancement from nanotwin formation.

11
 2.3.2. Planar Slip

While twinning may play a role in crack-deflection, other mechanisms have been
investigated, particularly to explain the differences in fatigue behavior of the same
material subject to different conditions. In addition to twins, Tang et al. observed strong
evidence of slip lines and bands after a sufficient number of cycles in regions of high
plastic deformation; these slip bands were described as potential sites of microcracking
and crack propagation within these regions [33]. Interesting, Niendorf et al. further
theorized that planarity slip may lead partially to the discovered partial deformation
reversibility observed in the scrutinized TWIP-HEA which in turn may contribute to high
fatigue limits [36].

2.3.3. Roughness-induced crack closure

Crack closure effects have further been suggested as the dominant mechanism
responsible for the downward shift of the fatigue curve with increasing R-ratio. In
metallic alloy systems, increasing load ratios often lead to reductions of fatigue
thresholds coupled with an increase in the Paris exponent. Seifi et al. postulated that R-
ratio dependency of HEAs reflected crack closure effects corresponding significant
surface roughening at low R-ratios [35]. The smoother, less-deformed fracture surfaces
would allow for greater physical contact of the crack flanks, leading to a reduction in the
crack-tip opening displacement, thus preventing plastic grain deformation in the crack
wake at cryogenic conditions resulting in increased fatigue-thresholds [37]. However, the
degree to which this change may be attributable to crack-closure is unclear and warrants
further investigation.

2.4. DISCUSSION

2.4.1. Importance of processing method

As shown by the work of Hemphill et al. and Tang et al., particular microstructural
characteristics of the alloy play a critical role on the overall performance of an HEA under
fatigue conditions. Scanning electron microscopy (SEM) imaging of the fracture surfaces
confirmed voids and/or inclusions were present in the fatigue-crack initiation region and
likely propagated between pores [33]. As the theory of the weakest link suggests, stress
concentrators and flaws (e.g. micro cracks, voids, impurities…) in the highly loaded
gauge area of a specimen can drastically reduce its fatigue lifetime, holding true for HEAs
both experimentally and computationally. In general, increased surface roughness due to
local plastic deformation in turn generates points of stress concentration that further

12
facilitate crack initiation. Local plastic deformation may be introduced through a variety
of mechanisms including but not limited to the effect of uneven thermal gradients during
processing (such as casting) and cold-working; however, annealing and heat-treatment
can alleviate its impact on the final specimen. This theory further explains the manner in
which scatter of fatigue tests can be significantly decreased by using high-purity
elements. In order to reach the theoretic upper bound endurance limits and further to
decrease the scatter of test results, continual improvement of the manufacturing process
is desirable [29,30]. In addition to preventing flaws, the tuning of other microstructural
characteristics, such as grain size, can also be conducted through modification of the
processing procedures to further improve the mechanical properties of these materials as
shown by ultrafine grained HEA studies of Liu et al [34].

2.4.2. Comparisons to other metallic alloy systems

The fatigue properties of the investigated HEA systems were found to largely
outperform most of the modern material compositions. This trend is particularly evident
in the total-life tests when using the metric of the ratio between fatigue threshold and
ultimate tensile stress for comparative purposes as shown in Figure 2.4. This comparison
method was suggested by Seifi et al. to provide a more rounded evaluation of this
mechanical property as materials demonstrating higher ultimate tensile stress often
exhibit lower ductility was well [15,33,35]. High entropy alloys seem to outclass common
metallic systems such as aluminum- and magnesium-based alloys and even some steels
in this respect. In comparing endurance limits of the various metals, Tang et al. found that
of the aluminum-variant HEA was elevated relative to steels and Ni-based superalloys
and was only rivaled by the bulk metallic glasses (BMGs) which demonstrate undesirable
brittleness compared to the ductility of HEAs; these results can be found in Figure 2.5
[33]. Depending on the processing method, the endurance limit of Al 0.5CoCrCuFeNi
ranged between a stress amplitude of 275 and 550 MPa, while the steels and superalloys
fell between 100 and 250 MPa depending on the alloy system.

13
Figure 2.4: Plot depicting the relationship between the fatigue ratios and the ultimate tensile stresses of
𝐴𝑙 . 𝐶𝑜𝐶𝑟𝐶𝑢𝐹𝑒𝑁𝑖 and other common engineering alloys. The HEA was found to exhibit a range of fatigue
behavior which was comparable to standard steels on the lower end of the spectrum while on the higher
end, low-defect HEAs outclassed even bulk metallic glasses [33].

Figure 2.5: Comparison of the fatigue endurance limits at 10 7 cycles between the 𝐴𝑙 . 𝐶𝑜𝐶𝑟𝐶𝑢𝐹𝑒𝑁𝑖 and
other conventional alloys. The HEA was found to outperform most of the common engineering materials
though also displayed wide scatter depending on the precise processing conditions [33].

14
 In the presence of a pre-existing flaw, the crack propagation behavior of the tested
HEAs was found to demonstrate behavior either very comparable or superior to other
commonly used structural and low-temperature steels depending on microstructure at
an R-ratio of 0.1. The as-cast Al-variant compositions Al 0.2CrFeNiTi0.2 and AlCrFeNi2Cu
showed comparable Paris slopes of ~5 and 3.5 but elevated fatigue thresholds of 16 and
17 MPa respectively, which may be due to the as-cast microstructure of these systems
[35]. While limited, the presented research suggests promising properties for possible
future applications of HEAs in which fatigue may occur as HEAs seem to be a safer
alternative for cutting-edge applications such as aerospace with excellent malleability
and ductility coupled with high endurance limits [9,36], but much more work must be
done to fill the large gaps of knowledge and understanding currently present in the
literature.

2.4.3. Potential directions for study

In general, the dominant conclusion reached upon surveying the current body of
literature is the lack of information available with ample room for further investigation
in nearly all respects from compositional and microstructural alterations to testing
conditions. Only a very small fraction of the growing HEA family has been tested, and of
those, the specimens have varied in processing, microstructure and test method, making
proper comparison between the various HEAs very difficult.

With regards to methodology, several conclusions regarding testing procedure going

forward have arisen from the aforementioned works. The work of Hemphill et al. and
Tang et al. stress the importance of sample preparation with high purity elements and
defect-removal for optimal and consistent performance with minimal scatter [31-33]. The
authors further discussed the importance of implementing tensile-fatigue testing – in
contrast to the performed four-point bend tests – to allow for ease of comparison of
fatigue test data with that of common engineering alloys and propose future work
focusing on this method. Niendorf et al. plan to further this research into the effects of
initially induced defects within the context of TRIP-HEAs and their performance under
high-cycle fatigue conditions [36].

Beyond alteration to specimen testing procedure, several specific avenues of research

have been proposed to help bridge some of the gaps between the works presented here
and to further explore the mechanisms underlaying the fatigue behavior of these alloys.
For example, Seifi et al. plan to further examine the temperature dependence of the two
phases Al0.2 CrFeNiTi0.2 and AlCrFeNi2Cu alloy systems [34,35]. However, they plan give
special attention to the bcc brittle-to-ductile transition at higher temperatures and the

15
resulting impact on the fatigue-crack growth rate and thus the potential for high-
temperature fatigue-related applications for HEAs rather than low temperature testing.

There also exists room for improvement outside of the laboratory within the
theoretical and computational realm, but the development of these sub-fields require
more empirical work to be conducted and to serve as the foundation for future
predictions. Preliminary computational models seem to be able to forecast tendencies of
the fatigue-behavior, but require additional low-scatter data for accurate results and are
currently only applicable to a severely small range of the HEA family; Hemphill et al.
applied a Weibull mixture model predicted an endurance limit of 858 MPa for the Al 0.5
CoCrCuFeNi alloy, while the test results found it to be between 540 and 945 MPa [31,32].
Tang et al. have further built on this work, having shown acceptable predictions of the
fatigue lifetime with a random endurance limit model, but to develop more precise
models, additional low-scatter test results are needed [33]. Additionally, the mechanisms
underlying the fatigue behavior of HEAs remain somewhat murky, hampering
understanding of the fundamental nature of this mechanical behavior. Future research is
needed to fully understand the microstructural role and impact on the fatigue behavior
of HEAs to utilize their outstanding mechanical behavior in modern applications.

16
 Chapter 3
Fatigue Characterization of the Cantor Alloy
Our society's demand for structural materials tailored to fit an increasingly wide
range of applications, coupled with the development of a vast array of processing
techniques, have spurred great interest in alloys that flout convention and may exhibit
hitherto unknown yet desirable properties. To help fill this role, high-entropy alloys and
the related ‘medium-entropy alloys’ (MEAs; comprising 3 or 4 elements in near-
equiatomic concentrations) are being examined in earnest as a class of seemingly
unlimited new structural materials based on the belief that many new alloys with useful
properties may be discovered [1-5,7-12,36-43]. One of the most prominent sub-classes of
HEA and MEA alloy families is that of the CrCoNi system. Indeed, the five-component
equiatomic CrMnFeCoNi alloy, first described by Cantor in 2004 [1], is a single-phase
face-centered cubic solid solution with remarkable mechanical properties. Specifically,
the alloy displays tensile strength levels of ~1 GPa associated with significant strain
hardening, excellent ductility (~60-70%) and exceptional fracture toughness (KJIc > 200
MPa√m) [4,6,8,36-38]; moreover, these properties tend to improve with decrease in
temperature between ambient and 77 K, i.e., they display damage-tolerant properties that
tend to run counter to the vast majority of metal alloys that show a propensity for
increasingly brittle behavior at cryogenic temperatures [4,8,11,44].

Whereas good strength and toughness are invariably vital characteristics for a
structural material, resistance to fatigue is often the limiting mechanical property, as this
generally determines the engineering lifetime of a given component for many
applications. As noted above, the mechanical properties of the CrMnFeCoNi alloy have
been well characterized under monotonic loading [4,7,8,10,11,36,38], but there is little
information to date how this alloy performs under cyclic loads. As a potential new class
of structural materials, it is clearly important that the fatigue behavior of HEAs and the
various factors that may affect it, such as grain size and temperature dependence, are
clearly understood. To date there have been only a few studies on the fatigue of HEAs,
but these have generally been focused on as-cast, rather than recrystallized materials with
uniform grain size; moreover, they have all been performed at room temperature
[31,35,45]. Thus I sought to characterize the high-cycle fatigue-crack propagation
behavior of the CrMnFeCoNi alloy, using material with a uniform recrystallized grain
size. Furthermore, as this alloy is particularly notable for its cryogenic damage-tolerant
properties, I performed these experiments at both room temperature and 198K, to
evaluate how the improved strength, ductility and toughness of the material at cryogenic
temperatures translates into its fatigue-crack growth resistance.

17
3.1. METHODS AND MATERIALS

An ingot of the CrMnFeCoNi alloy was produced at the Ruhr-University, Bochum,

through vacuum induction melting and casting into a cylindrical mold. The ingot was
then sealed in a quartz tube and thermally homogenized at 1473 K for 48 hr.
Subsequently, the homogenized rod was rotary swaged at room temperature to reduce
its diameter from 40 to ~16.5 mm before being recrystallized at 1073 K for 1 h, yielding a
grain size of 7 ± 3 µm with a random orientation distribution [46]. The rod was then sliced
into disc-shaped compact-tension (DC(T)) samples, in accordance with ASTM standard
E1820 [47], using electrical discharge machining. 27 samples (N = 27) were machined with
a width, W of 12.5 mm and thickness, B of 6 mm; notches were cut with depths varying
from a0 = 3.6 to 5.1 mm, corresponding to initial a0/W ratios of ~0.28 to 0.4, respectively,
and with notch root radii of ~100 µm.

The faces of all samples were metallographically polished using silicon carbide paper
to a final 1 µm surface finish to allow accurate crack-length measurements on the surface
of the samples using optical microscopy techniques. To allow continuous monitoring of
the crack growth rate during testing, a linear patterned strain gauge (Vishay Precision
Group, Raleigh, NC, USA) was applied to the back-face of each specimen: model EA-06-
031DE-350 gauges were used for tests at 293 K or 198 K and model WK-13-031DE-350
gauges at 77 K, the latter to permit accurate readings in the deep-cryogen environment.
The strain gauges were wired into a Wheatstone bridge to amply the signal for accurate
monitoring. Crack lengths were calculated from the strain gauge readings during the
elastic unloading portion of each cycle using the compliance expression for the DC(T)
sample with back-face strain, as described by Ritchie et al. [48]:

a/W = 0.796239 + 5.40205u - 103.821u2 + 714.676u3 - 2603.44u4 + 4829.01u5 - 3578.51u6 , (1)

where u= . (2)
√

Here E corresponds to the Young's modulus of the material, and C represents the
compliance calculated as the reciprocal from the unloading slope of the samples during
testing. This calibration is considered valid for the range 0.3 ≤ a/W ≤ 0.8; all samples were
within this range for the entire duration of testing to ensure measurement validity.

Before testing, all samples were fatigue pre-cracked in ambient conditions using an
electro-servo hydraulic MTS testing machine (MTS Corporation, Eden Prairie, MN, USA)
controlled by an Instron 8800 digital controller (Instron Corporation, Norwood, MA,
USA). Fatigue pre-cracking was conducted under load control (tension-tension loading)

18
within a stress-intensity range, ΔK = Kmax - Kmin, between 8 to 10 MPa√m at a constant
sinusoidal frequency, f, of 25 Hz and a load ratio, R = 0.1, where R is the ratio of minimum
to maximum load. For tests conducted at higher R, further pre-cracking was conducted,
increasing the load ratio of pre-cracking in a stepwise manner until the testing R was
matched in order to avoid excessive plasticity effects at the start of testing. The back-face
strain gauges were calibrated and balanced at zero-load to ensure residual strain from
mounting would not affect crack-length measurement, which was further verified during
pre-cracking by optically checking both sides of the samples for crack length and
linearity. Final overall pre-crack lengths including notches ranged from a = 3.7 mm to 5.17
mm (a/W ~ 0.3-0.41), all in accordance with ASTM 647 [49,50]. Images of components of
the experimental set-up may be viewed in Figure 3.1.

Figure 3.1: Images of sample preparation and mounting in the Instron load frame. Clockwise from left:
Wheatstone bridge consisting of three 350 Ω for signal amplification, sample with attached strain gauge
and full wiring, side view of prepared samples, and sample mounted in load frame.

Fatigue-crack growth tests were performed under sinusoidal cyclic loading at f = 25

Hz at one of three constant load ratios of R = 0.1, 0.4, and 0.7 in three different temperature
environments: room temperature air (293 K), dry ice/ethanol bath (198 K), and liquid
nitrogen (LN2) (77 K); thus yielding a total of 9 testing condition sets. All fatigue-crack
propagation testing was performed under load-shedding or constant load control
conditions. Near-threshold crack-growth rates were determined under load-shedding
conditions, which was automated to decrease the load at a rate such that the normalized
19
K-gradient remains above -0.08 mm-1, as recommended in ASTM E647 [51]. ΔKTH fatigue
threshold stress-intensity values were determined as the maximum value of the stress-
intensity range, ΔK, to give a growth rate approaching 10-11 m/cycle. Higher regions of
the crack-growth curves were characterized under constant load conditions, in which the
amplitude of the sine wave is held constant, effectively increasing the stress-intensity
range as the crack grows, while the rest of the fatigue curve was determined by an overlap
of the two methods. To obtain the moderate cryogenic test conditions, a dry ice/ethanol
bath was prepared and maintained at 198 K for the duration of the test. To obtain the low
cryogenic test conditions, a LN2 bath was continually maintained to keep the sample
submerged throughout testing. Samples subjected to 198 K and 77 K conditions were
tested in a continuous manner at this temperature to avoid potential thermal cycling
effects [52]. For all samples, crack length verification by optical microscopy was carried
out after testing after samples had been brought back up to room temperatures in an
ambient temperature ethanol bath to avoid oxidation of the fracture surfaces.

To determine the microstructural mechanisms controlling the propagation of the

fatigue cracks in this alloy and to investigate the nature of the crack paths (Chapter 4),
select samples were sliced longitudinally through the center for back-scattered electron
(BSE) and electron backscatter diffraction (EBSD) analyses, performed using a Zeiss LEO
1525 field-emission scanning electron microscope (FE-SEM, Carl Zeiss, Oberkochen,
Germany) operated at 20 kV using a TEAMTM EDAX analysis system (Ametek EDAX,
Mahwah, NJ, USA). To perform fractography analyses, samples from both temperatures
were subsequently quasi-statically loaded to complete failure to have unrestricted access
to their fracture surfaces. In addition to the microscopic analysis with the Zeiss-SEM,
fracture surface characterization was also performed using a JSM-7500F SEM (JEOL USA,
Arvada, CO, USA) operated at 5-15 kV in secondary electron (SE) mode. Further BSE and
EBSD imaging was conducted with an FEI Strata DB235 SEM (FEI Company, Hillsboro,
OR, USA) operated at 20 kV using a TEAMTM EDAX analysis system (Ametek EDAX,
Mahwah, NJ, USA) with step sizes of 0.12-0.1 µm. Post-processing of these EBSD images
was conducted using OIM Analysis v6 software (Ametek EDAX, Mahwah, NJ, USA).

3.2. RESULTS

3.2.1. Fatigue testing at 293 K and 198 K with R = 0.1

Prior studies conducted on the equiatomic, single-phase fcc CrMnFeCoNi alloy with
the same microstructure have found it to display mechanical properties that either remain
consistent or improve as the testing temperature is lowered into the cryogenic regime. In
this regard, both strength and ductility of the material are particularly notable for their

20
strong temperature dependence; the transition from 293 K to 198 K yields an increase in
ultimate tensile strength from ~760 MPa to ~925 MPa with a similar increase in yield
strength from ~410 MPa to ~520 MPa while its ductility increases from ~0.6 to ~0.7 [8].
Over the same temperature range, the Young’s modulus [53,54], toughness and strain
hardening exponent remain relatively unchanged. While both Young’s modulus, E, and
fracture toughness, KJIc, increase slightly from 202 GPa to 209 GPa and from 217 MPa√m
to 221 MPa√m, respectively, the strain hardening exponent of the material remains
consistently high at ~0.4 [8] These trends, displayed by strength, modulus, toughness and
strain hardening coefficient measurements, have been found to continue as the
temperature is further decreased to 77 K.

Figure 3.2: Fatigue-crack growth behavior of the CrMnFeCoNi alloy, tested at a load ratio R of 0.1 at 293 K
and 198 K. A direct comparison of the fatigue-crack growth curves, as a function of the stress-intensity
range, ΔK, at both temperatures tested on disc-shaped compact-tension (DC(T)) samples at a frequency of
f = 25 Hz reveal a clear shift in the fatigue threshold, increasing some 30%, from ~4.8 MPa√m to ~6.3 MPa√m,
as the temperature was reduced from ambient to cryogenic conditions. The Paris exponent m was found to
change from 3.5 at 293 K to 4.5 at 198K. The smaller sized, solid symbols indicate the corresponding
variation, as a function of ΔK, in local crack-growth rates, which were estimated from striation spacing
measurements on the scanning electron microscopy images of the fracture surfaces.

21
 Here, the fatigue-crack growth rates, da/dN, of the alloy, for tests at 293 K in air and at
198 K in dry ice/ethanol, are plotted in Fig. 3.2 as a function of the applied stress-intensity
range, ΔK. Although crack-propagation rates are comparable in the mid-range of growth
rates above ~10-8 m/cycle with only a modest change of the Paris exponent m from 3.5 at
293K to 4.5 at 198 K, near-threshold growth rates and the value of the fatigue threshold,
ΔKTH, of this alloy are markedly improved at cryogenic temperatures. Specifically, growth
rates are lowered by up to an order of magnitude, and threshold ΔKTH values are
increased by ~30%, from ~4.8 MPa√m to ~6.3 MPa√m, with decrease in temperature from
293 K to 198 K.

At higher growth rates, traditional fatigue striations are visible on the fatigue fracture
surfaces of this high-entropy alloy (Fig. 3.3). These features are rather rounded and
characteristic of ductile striations. To analyze these features, I screened the fracture
surfaces of four samples specifically for fatigue striations, two each from the 198 K and
293 K crack-growth experiments. Sampling was conducted at intervals of 1 mm,
beginning 1 mm from the notch along the width of the sample until the overload region
along the centerline of the sample, with multiple images taken at each location to check
for consistency. While the fatigue striations at growth rates near the threshold are hard
to detect and cannot often be easily distinguished from slip steps, e.g., careful
investigation of the fatigue striations, as shown on the example of a sample tested at ΔK
~ 21.5 MPa√m (Fig. 3.3) shows good agreement (Fig. 3.2) between their spacing and the
macroscopic growth rate of ~10-7 m/cycle for both temperatures, as shown in Fig. 3.2. Fig.
3.2 also clearly indicates the actual correspondence of striation spacing as a function of
ΔK of the local crack-growth rates and the macroscopic growth rates measured at both
temperatures. At lower values of ΔK, the width of fatigue striations begins to diverge
from the macroscopic growth rates, behavior which is indicative of the non-uniform crack
propagation along the crack front at low growth rates as the crack advances in an
incremental fashion. Overall, striation spacing measurements for both sets of samples
displayed the generally linear trend characteristic of the approximate first-order
description of macroscopic crack-growth rate behavior, albeit with scatter due to
variability in measurements.

22
Figure 3.3: High magnification SEM of fracture surfaces at comparable ΔK~20 MPa√m. (a),(b) depict 198 K
tested sample in the region ΔK ~ 22.7 MPa√m while (c),(d) depict the ambient 293 K tested sample in the
region ΔK ~ 21.6 MPa√m to allow comparison between the two sets of images under similar loading
conditions. At this ΔK, the measured crack growth rate was ~ 1x10-7 m/cycle for both temperature regimes,
so we would expect fatigue striations to have a width of ~100 nm. This value corresponded well with the
observed striation rate in both sets of samples, with average widths of ~8x10 -8 m and ~9x10-8 for 198 K and
293 K respectively While the striation size was similar, the shape differed drastically between the two
temperatures. While samples tested at 198 K displayed sharp, angular striations on flatter surfaces, the
samples tested at 293 K show far more irregular, rounded striations.

3.2.1. Fatigue testing at R = 0.4 and 0.7

Upon the conclusion of testing at R = 0.1, trials at R = 0.4 and 0.7 commenced.
Additionally, the temperature range of testing was extended down to 77 K with sampled
immersion in cryogenic LN2. However, due to testing limitations, LN2 trials were limited
to testing under constant load conditions, and thus the linear portion of the Paris regime;
no threshold data was collected at this temperature. The summarized data may be viewed
in Figure 3.4 and Table 3.1

23
Figure 3.4: Fatigue crack growth rates as a function of the stress-intensity range for the Cantor alloy under
various testing conditions. Testing was conducted over three temperature ranges between ambient (298 K)
and liquid-nitrogen temperature (77 K) at R = 0.1, 0.4, and 0.7 at 25 Hz. Tests at 77 K were limited to the
upper region of the mid-growth rate regime while testing at 293 K and 198 K continued until the fatigue
threshold stress intensity was reached. A direct comparison of the crack-propagation curves characterizing
these various testing parameters indicates a clear upwards shift in the value of the KTH threshold as both
temperature and load ratio were decreased.

24
Table 3.1: Summary of fatigue-crack growth and threshold data, as shown in Fig. 3.4

Temperature Load Ratio (R) Paris Law exponent (m) Threshold (ΔKTH) [MPa√m]

0.1 3.5 4.8

293 K 0.4 2.8 2.6
0.7 2.6 2.5
0.1 4.5 6.3
198 K 0.4 3.4 4.8
0.7 3.1 3.5
0.1 2.9
77 K 0.4 2.3
0.7 2.3

As indicated by the data, for the mid-range of growth rates above 10 -9 m/cycle, the
Paris law exponents for all temperatures and load ratios were largely in the characteristic
range of m ~ 2 to 4, although in general there was a trend of increasing fatigue-crack
growth resistance, i.e., lower growth rates, at a given ΔK, and higher values of the fatigue
ΔKTH thresholds, at the lower temperatures. Indeed, ΔKTH values, which ranged from 6.3
to 2.5 MPa√m at load ratios between 0.1 and 0.7 were, respectively, 31% and 12% higher
at 198 K as compared to 293. Additionally, there was a marked effect of load ratio on
crack-growth rates and ΔKTH threshold values with increasing load ratios which became
more prominent at near-threshold levels. Specifically, ΔKTH threshold values are 2-3 times
lower at R = 0.7 compared to R = 0.1. Such load-ratio dependent behavior is typical of
most metallic alloys [51] and can generally be related to crack-tip shielding from fatigue
crack closure effects [55], as further discussed in Chapter 4.

Upon investigation of the fracture surfaces, traditional striation features could be

observed at R = 0.4 and R = 0.7 samples, with particularly clear visibility at high crack
growth rates. As with the striations observed in R = 0.1 tested samples, the striation width
of these striations generally agreed with the predicted crack growth rate in these regions.
Examples of observed striations of samples tested at 198 K are shown in Figure 3.5. While
visible at all values of R, increasing R seemed to correlate with increasing prominence of
striations which may be due to the greater plastic deformation of the sample with higher
load ratio. As with the R = 0.1 samples, as ΔK decreased, the striation width began to
deviate from the calculated crack growth rate and to become obscured by other surface
features, further discussed in Chapter 4.

25
Figure 3.5: Example of fatigue striations observed in the high growth rate region of samples tested at 198
K. (a) and (b) depict samples fatigued at R = 0.4 and 0.7 respectively. Each image depicts striations of width
on the order of the expected fatigue crack growth rate.

While these results satisfactorily summarize the basic fatigue data of the Cantor alloy
at ambient and low temperature ranges, this information does not explain ‘how’ or ‘why’
the Cantor alloy may demonstrate this temperature dependent shift in fatigue behavior.
As such, mechanistic studies were called for to explain these shifts and thus to potentially
shed light on future alloy design.

26
 Chapter 4
Mechanistic Studies of Fatigue in the Cantor Alloy
4.1. THRESHOLD ANAYSIS

The most prevalent reason for the marked effect of load ratio on fatigue-crack growth
rates, especially at near-threshold levels, is that of crack closure. As this phenomenon is
dominated by shielding effects due to wedging effects between the crack flanks, e.g., from
corrosion debris (oxide-induced crack closure) and/or from crack surface asperities
(roughness-induced crack closure), which act to reduce the effective stress-intensity
range actually experienced at the crack tip, crack closure is exacerbated at low ΔK levels,
where the cyclic crack-tip opening displacements (CTODs) are small and become
comparable in dimension to the thickness of the “wedge”, yet minimized at high positive
load ratios where the minimum opening displacements are much larger, specifically
where Kmin > Kcl. On the simple assumption that the value of the closure stress intensity
Kcl, and the effective stress intensity range at the threshold ΔKeff,TH are nominally constant
and independent of R, we would expect the measured variation in ΔKTH thresholds to
follow “Schmidt and Paris type” behavior [59]. Specifically, below a critical load ratio
where Kcl = Kmin (usually on the order of R ~ 0.4-0.5), the ΔKTH threshold should decrease
with increasing R whereas the threshold Kmax should remain roughly constant (this is the
regime where closure is active as Kcl > Kmin); above this critical R value though (where
closure is inactive as Kcl < Kmin), the value of ΔKTH threshold should reach a lower-bound
“plateau whereas the threshold Kmax should increase as R tends to unity. A plot of ΔKTH
and Kmax vs. R may be viewed in Fig. 4.1; although I have only tested three different load
ratios at each temperature, these measurements are still useful to determining if the
Cantor alloy’s fatigue behavior roughly follows these crack closure related trends
described by Schmidt and Paris.

Shown in Fig 4.1, the threshold data for this high-entropy alloy are not inconsistent
with this hypothesis. As previously discussed, lowered temperature was found to have a
pronounced impact on the stress intensity range at which a ‘threshold’ point of arrested
crack growth was reached, particularly evident in samples tested at R = 0.4. Samples
tested under ambient conditions at 293 K exhibited the aforementioned ‘plateau’ effect at
load ratios of R ≥ 0.4, staying roughly constant at ΔKTH ≈ 2.5 MPa√m. However, this
leveling-off behavior was not seen under cryogenic conditions among the R-ratios tested,
with a continued decrease evident past R = 0.4. Following Schmidt and Paris, this would
in turn indicate an upwards shift in the critical load ratio at which Kcl = Kmin, though due
to the small sampling of R-ratios, this point can only be said to be at R > 0.4.

27
Figure 4.1: Plot of threshold ΔKTH and KTH,MAX values of samples tested at 293 K and 198 K with respect to
R-ratio

However, pertinent questions that remain are whether this closure affect can be
associated with fracture surface roughness, which is the most prevalent form of fatigue
crack closure in metallic materials, whether such phenomena can explain the observed
increase in ΔKTH thresholds in this HEA at cryogenic temperatures, and whether
deformation twinning, which is so prevalent during overload fracture in this alloy at 77
K [8,10] plays an active role during fatigue-crack growth.

4.2. CRACK CLOSURE STUDIES

While the Schmidt and Paris approach provides a qualitative indicator that crack
closure may be a contributing factor to the shift in threshold with temperature, a more
quantitative approach was required. In order to calculate the degree to which crack-
closure may be an active mechanism in near-threshold fatigue testing in the Cantor alloy,
select near-threshold tested specimens were further subjected to limited loading and

28
unloading cycles to estimate the extent of crack closure from measurement of the stress
intensity for crack closure, Kcl. This was used to estimate an effective stress-intensity
range actually experienced at the crack tip, where ΔKeff = Kmax - Kcl, where Kcl > Kmin (where
Kcl < Kmin, ΔKeff = ΔK). Full details of these procedures can be found elsewhere [55,60]. These
loading cycles were performed in incremental steps corresponding to an increase in ΔK
of 0.5 MPa√m beginning at the threshold of the respective sample and the resultant
position-displacement curves were plotted and analyzed for the distinctive ‘kink’ that
indicates the popping open of the crack tip and thus Kcl for the analyzed region; this
analysis, only the threshold region of the curve was examined. Additionally,
approximations for the crack tip opening displacement (CTOD) at each threshold were
calculated to compare with the approximate asperity sizes in the region. These results can
be viewed in Table 4.1. Due to the lack of threshold data for 77 K, this analysis was limited
to 293 K and 198 K tested samples.

Table 4.1: Summary of fatigue crack growth data with accompanying crack closure data
Load Paris Law Threshold (ΔKTH) ΔCTOD CTODmax Crack Closure ΔKTH,eff
Temp.
Ratio (R) exponent (m) [MPa√m] [nm] [nm] (Kcl) [MPa√m] [MPa√m]

0.1 3.5 4.8 130 160 1.38 3.95

293 K 0.4 2.8 2.6 40 105 - 2.6
0.7 2.6 2.5 35 390 - 2.5
0.1 4.5 6.3 170 210 2.15 4.85
198 K 0.4 3.4 4.8 100 275 2.7 - 3.4 4.6 - 5.3
0.7 3.1 3.5 55 585 - 3.5
0.1 2.9
77 K 0.4 2.3
0.7 2.3

As summarized in Table 4.1, crack closure levels are far higher at low load ratios as
expected. Moreover, the effective ΔKeff,TH thresholds appear to be roughly constant, again
consistent with our observations of the marked effect of load ratio on the near-threshold
fatigue behavior being associated with a crack closure effect. With regards to temperature
trends, the data validates the prediction from the Schmidt and Paris analysis that the
critical R-ratio for which Kcl = Kmin has shifted to a higher load ratio from R ≤ 4 at 293 K to
4 ≤ R≤7 at 198 K, as crack closure is still observed at R = 0.4 but not at R = 0.7 in samples
tested at 198 K. In both temperature ranges, crack closure appears to play a major role in
the near-threshold fatigue behavior of this alloy.

29
4.3. CRACK PATH STUDIES

Examination of the crack paths was performed by sectioning samples at mid-

thickness. At R = 0.1, samples tested at room temperature showed crack path behavior
was primarily transgranular, particularly at high growth rates, with some minor
intergranular crack growth at low and intermediate growth rates, as is apparent from the
BSE images and EBSD scans in Fig. 4.2. The BSE images of the crack-tip region near the
threshold at ΔK = 4.8 MPa√m clearly indicate that the propagating crack runs through the
grains and recrystallization twins and not along their boundaries. EBSD images taken
from the wake of the crack at ΔK values of ~6.0 MPa√m and ~7.0 MPa√m show significant
amounts of plastic deformation in the grains neighboring the crack flanks partly resulting
from geometrically-necessary dislocations emitted from the crack tip during loading and
unloading of the samples. Additionally, such plastically-deformed regions naturally
occur as a result of local deviations from a straight crack path leading to physical contact
of the mating crack surfaces resulting from small variations in the mode II displacements
of the neighboring crack flanks. This is known as roughness-induced fatigue crack
closure and acts to lower the crack-driving force experienced at the tip of the growing
crack by effectively increasing the value of the minimum stress intensity in the fatigue
cycle [56,57].

At room temperature, where the alloy has a lower strength compared to at 198 K, such
closure can result in local plastically deformed regions near the propagated crack and
particularly of the individual grains next to the crack flanks. This mechanism can to some
extent account for the difference found in the fatigue threshold shown in Fig. 3.2.
Interestingly, despite the relatively large stresses near the crack tip, deformation induced
nano-scale twinning could not be observed by BSE and EBSD microscopy within these
room temperature tested samples.

30
Figure 4.2: Crack path characterization of the CrMnFeCoNi alloy fatigue tested at 293 K and R = 0.1. Back-
scattered electron (BSE) images and electron back-scatter diffraction (EBSD) scans taken at the mid-
thickness of the samples near the threshold at ΔK = 4.8 MPa√m, specifically from the wake of the crack at
ΔK values of ~6 and ~7 MPa√m, clearly show transgranular fracture as the dominant crack propagation
mode. Images taken at the crack tip indicate that the crack propagated through the grains and
recrystallization twins and not along their boundaries. EBSD scans taken from the wake of the propagating
crack show plastically deformed region along the crack flanks resulting from physical contact between the
mating crack surfaces and the lower strength of the alloy at room temperature.

Figure 4.3 displays the crack paths of the fatigue cracks at tested at R = 0.1 and in the
198 K temperature regime. Note that akin to the behavior at 293 K, we could not see
evidence of nanoscale deformation twinning associated with the crack-growth
mechanisms in our BSE and EBSD analyses, although as noted below, there appears to be
evidence of the onset of twinning on the fracture surfaces. However, compared to the
room-temperature behavior, both BSE images and EBSD scans clearly show significant
amounts of intergranular fracture at this lower temperature; like at room temperature
though, the presence of recrystallization twins did not appear to influence the crack
trajectory. As a result of the intergranular crack propagation at the lower temperatures,
the local crack path was naturally more deviated compared to that at room temperature,
thereby giving rise to significantly more contact of the neighboring crack flanks to
promote roughness-induced crack closure; this is consistent with the 30% higher
threshold at 198 K, ΔK = 6.3 MPa√m, as compared to ΔK = 4.8 MPa√m at 293K; this crack
closure mechanism operates particularly effectively within low stress-intensity range
conditions like those of the near-threshold region [57,58]. Additionally, the higher
strength at the lower temperature acts to limit the crack-tip opening displacements at a

31
given ΔK, as this displacement is proportional to ΔK2/σyE, where σy is the flow strength
and E the Young’s modulus; correspondingly, the generation of new crack surface during
the opening of the crack in each cycle will be less effective.

Figure 4.3: Crack path characterization of the CrMnFeCoNi alloy fatigue tested at 198 K and R = 0.1. In
contrast to the behavior at room temperature (Fig. 4.2), both BSE images and EBSD scans revealed
intergranular fracture as the dominant crack propagation mechanism at 198 K. The boundaries of the
recrystallization twins do not appear to influence the path of the crack tip, as can be seen near the threshold
at ΔK = 6.3 MPa√m and at higher growth rates at ΔK values of ~7.5 MPa√m and ~8.5 MPa√m; this is believed
to be a result of the higher strength of the material at 198K, such that plastically deformed regions along
the propagated crack do not to occur.

Together, these two sets of images lend strong evidence to a lack of twinning as a
dominant mechanism in the fatigue behavior of these alloys; should deformation
twinning be a factor in the temperature dependence of fatigue in this alloy, then their
presence would have been noted along the crack path of samples tested at 198 K. Rather
than deformation twins, potential regions of crack closure were noted, further indicating
the likely dominance of this mechanism in the fatigue threshold shift. However, while
deformation twinning does not likely play a role at the threshold, it does appear to
become activated at sufficiently high ΔK (or, rather, Kmax) and sufficiently low
temperature. As displayed in Fig. 4.4, the samples tested in LN2 at R= 0.1 and R = 0.7
subjected to high ΔK were both found to contain evidence of a small number of
deformation twins that formed in grains along the crack path.

32
Figure 4.4: EBSD imaging of the crack path in the high stress intensity regions of LN2 tested samples. To
examine the potential role of nano-twinning in the shift of the fatigue curve in cryogenically-tested
CrMnFeCoNi, detailed EBSD scans were taken at the crack-tip region, the area at which the largest stress
intensities were experienced. Images (a) and (b) correspond to imaging of samples subjected to R = 0.1 with
ΔK ≈ 29 MPa√m and R = 0.7 with ΔK ≈ 20 MPa√m respectively with crack propagation direction indicated
by inset arrow. Image (c) shows a higher resolution portion of the crack path of the R = 0.7 sample as
indicated by the inset highlight box.

33
4.4. FRACTOGRAPHY

As can be expected from the BSE images and EBSD scans shown in Fig. 4.2, SEM
characterization of the fracture surfaces of samples fatigued at 293 K and R = 0.1 revealed
mainly transgranular crack propagation (Fig. 4.5 a-c) with some minor intergranular
failure regions (inset of Fig. 4.5 c). Most of the transgranular failure region is covered with
highly serrated fracture surface features at both the threshold (Fig. 4.5 a) and at higher
growth rates (Fig. 4.5 b,c). The serrations on the fracture surfaces are clearly different in
size and shape than typical striations which were found at higher growth rates, as
described below; they have far sharper edges (insets Fig. 4.5 a-c) and appear to form at
specific angles with respect to each other (insets Fig.4 a,c). Given that at room
temperature, nanoscale deformation twinning has not been found as dominant
deformation mechanism in this alloy in both strength and fracture toughness tests [8,10],
the latter having significantly higher distributions of crack-tip stresses than those
generated during fatigue-crack growth, these periodic serrations cannot be attributed to
twinning as the prominent deformation mechanism occurring at the crack tip. Instead,
they can only be associated with cyclic slip steps resulting from dislocation motion by
planar slip which has been confirmed as the main deformation mechanism for the
material at this temperature [10].

Similar to the fracture morphology at 293 K, the fracture surface in the near-threshold
region of R = 0.1 samples tested at 198 K, at ΔK = 6.3 MPa√m, is highly serrated with sharp
edges that form at specific angles, as shown in Fig. 4.5 d, indicating that planar slip is also
a major deformation mechanism at this temperature. Crack propagation at higher growth
rates occurred mainly through intergranular failure, as shown in Fig. 4.5 e,f and the inset
of Fig. 4.5 f, consistent with the BSE images and EBSD scans shown in Fig. 4.3.
Interestingly, polyhedral features can be found on numerous grain boundaries, as
indicated by the white arrow in Fig. 4.5 d and the insets of Fig. 4.5e; we believe that these
features are the ends of deformation twins formed perpendicular to the grain boundaries
into the grains neighboring the crack front. The increased occurrence of these features at
198 K is consistent with the higher strength of the alloy at these lower temperatures as it
is known that deformation twinning in this alloy is a stress-controlled phenomenon [46].
In addition, the serrated straight lines on the grain boundaries appear to be twins that
have formed parallel to these grain boundaries rather than perpendicularly to them
thereby indicating that twinning at the lower temperature seems to occur more readily
than it does at room temperature. This mechanism is consistent with the differences in
both the fatigue threshold and the growth rates in the near-threshold regime between 293
K and 198 K.

34
Figure 4.5: Fractographic analysis of CrMnFeCoNi samples tested at R = 0.1 in the 293 K and 198 K
temperature regimes. Fracture surface characterization of samples tested at 293 K show mainly
intergranular crack propagation (a-c) with some minor transgranular failure regions (inset of (c)). The
highly serrated fracture surface features covering most of the intergranular failure region (insets of (b) and
(c) are likely associated with cyclic slip steps resulting from dislocation motion by planar slip [10]. Similarly,
the near-threshold region at 198 K (d) shows planar slip as major deformation mechanism at this
temperature whereas at higher growth rates (e,f) crack propagation occurs mainly through intergranular
fracture. The white arrow in (d) and the insets of (e) indicate polyhedral features that likely show the ends
of deformation twins formed perpendicular to the grain boundaries into the grains neighboring the crack
front.

35
 In general, the same trends were largely observed to continue even with testing at
higher values of R. Regions of the samples tested at higher ΔK tended to display rougher
surface features with prominently displaced striations, a result of the increasing plasticity
with increasing stress intensity range applied to the surface. Regions of the samples that
were subject to lower values of ΔK, by contrast, tended to show a greater prominence of
facet-like features likely characteristic of intergranular fracture. One example of this
behavior at higher R may be viewed in Fig. 4.6, which depicts the fracture surface of a
sample tested at 198 K at R= 0.4 in at both moderate ΔK ~ 9 MPa√m and a higher ΔK ~ 17
MPa√m. Note both images are taken at or beyond the higher end of the range shown in
Fig. 4.5.

Figure 4.6: Fractographic analysis of CrMnFeCoNi samples tested at R = 0.4 at 198 K at two regions subject
to different stress intensity ranges. (a) depicts a region subject to moderately high ΔK ~ 9 MPa√m while (b)
shows a region subject to high ΔK ~ 17 MPa√m.

Direct comparison between fracture surfaces as a function of R-ratio was also

attempted between all nine testing conditions, as displayed in Fig. 4.7. This figure
portrays regions of each sample type subject to a stress intensity range of approximately
ΔK ~ 12 MPa√m; as such, the predicted crack growth rate depicted varied as per Fig. 3.4.
Unfortunately, due to a degree of post-testing surface oxidation from thermal cycling,
fractography of the three LN2 samples proved inconclusive; this oxidation likely occurred
due to the condensation of water vapor on the surfaces as samples were removed from
the non-reactive nitrogen atmosphere prior to rewarming to ambient temperature.
Images from samples tested at 198 K and 293 K indicate a general trend towards increased
surface roughness and striation prominence as R-ratio increases. This phenomenon may
be in part due to the increased CTOD observed with higher R-ratio: due to the increased
opening of the crack tip (and accompanying decrease and elimination of crack closure),
fracture surfaces of samples tested at high R-ratio are not subject to the same degree of

36
wear from the cyclic closing and opening of the fracture surfaces that those testing at
lower R may undergo and thus would tend to show sharper features. In general, the
fracture surfaces continue the trends previously observed in samples tested at R= 0.1,
with a transition from rougher, more transgranular fracture behavior to smoother,
angular, and more faceting intergranular behavior with decreasing temperature and
decreasing R.

Figure 4.7: Overview of fractography over all nine testing conditions at a moderately high ΔK ~ 12 MPa√m.
Images of fracture surfaces at 77 K (left column) show evidence of post-testing oxidation though the general
underlying shapes correlate well with the trends observed between 198 K and 293 K tested samples. Higher
R seemed to correlate with increased surface plasticity and sharpness of features, as predicted.
Additionally, a greater degree of flatter, facet-like features with lower temperature continued to be
observed at this higher ΔK though not to the same degree at in Fig. 4.5, as expected by the observed
increasing degree of transgranular fracture behavior with higher ΔK.

37
4.4. SUMMARY

Overall, the mechanistic studies performed lend great credence to the hypothesis that
the dominant mechanism involved in the shift of the fatigue threshold in the Cantor alloy
is that of crack-closure, specifically roughness-induced crack closure. These conclusions
are supported by the observations of each section outlined above. Schmidt and Paris
analysis allowed an initial, more qualitative check for the potential presence of crack
closure, while the quantitative crack closure studies found definitive evidence of crack
closure and its different impact at different temperatures. Crack closure could only be
measured in samples tested in ambient conditions in samples tested at low R-Ratio, (R =
0.1), while it could be observed in samples tested at cryogenic 198 K for an expanded
range of R-ratio, up to R = 0.4.

BSE, SE, and EBSD studies of the crack path and fracture surfaces provide visual
support to these numeric calculations. Little evidence of deformation twinning, a
potential competing mechanism, was observed along the crack path until the imposition
of very high stress intensities at very low temperatures (77 K), while some evidence of
potential crack closure was observed. Additionally, both crack path studies and
fractography indicated a transition from rougher, transgranular fracture to smoother,
more intergranular character with the lowering of temperature and R; these surfaces
would allow for the ‘rezipping’ effect of crack closure more readily than the more
plastically warped surfaces. Indeed, the smoother surfaces at lower R may also indicate
the occurrence of crack closure in and of themselves, as repeated contact of the two
surfaces may lead to a wearing effect, thus causing a smoothening of surface features.
Thus, I conclude crack closure plays an important if not dominant role in the temperature
and R-ratio dependent fatigue behavior of the Cantor alloy.

38
 Chapter 5
Discussion

5.1. COMPARISON TO OTHER ALLOY SYSTEMS

Based on the previously reported mechanical properties of the CrMnFeCoNi alloys in

combination with its fatigue performance shown here, the alloy demonstrates high
potential as a structural material for engineering applications, especially at lower
temperatures [8].

Compared to other HEAs, austenitic stainless steels and twinning-induced plasticity

(TWIP) steels, all tested at 298 K at similar testing conditions and comparable in terms of
microstructure, the Cantor alloy shows excellent fatigue crack-growth performance at
both 298 K and 198 K, as shown in Fig. 5.1; details about the individual materials and
testing conditions can be found in Table 5.1. The fatigue-crack growth behavior of the
CrMnFeCoNi alloy bore the most similarity to that of TWIP steels [61,62]. Akin to our
material, TWIP steels also have displayed a lack of twinning behavior under fatigue
conditions at room temperature [61,62]. However, they are also known for their high
strength complemented with large ductility attributed to their propensity towards
deformation twinning at lower temperatures [66]. As shown in Fig. 5.1, both their fatigue
thresholds and their Paris slopes are comparable to our materials. Furthermore, the
CrMnFeCoNi alloy compares favorably in terms of its fatigue threshold to several widely
used high-strength steels of both low and moderate carbon content with comparable
strength although crack-growth rates in the Paris regime are invariably fairly similar [63-
65].

In contrast, the Cantor alloy displayed a lower fatigue threshold to the related Al-
containing HEAs [35], but the reason for the disparity may be due to the differences in
microstructure between the alloys, particularly between the as-cast structure of the Al-
based HEAs and the uniformity of our ~7 µm grain size CrMnFeCoNi, rather than due to
inherent compositional differences. As-cast structures have been found to exhibit higher
fatigue thresholds than their more finely grained counterparts, which is linked to an
increased fracture surface roughness [67]. Moreover, the initially published fatigue data
[35] on the Al-CrFeNi-based HEAs may be in error, in terms of unacceptably high fatigue
threshold values above 20 MPa√m, due to the small size of the specimens tested and the
correspondingly excess plasticity throughout their cross-section. Data on these Al-

39
containing HEA using larger specimen size, in comparison to the extent of local plasticity,
should resolve this issue.

Figure 5.1: Comparison of the fatigue-crack growth behavior of the CrMnFeCoNi high-entropy alloy with
austenitic stainless steels and TWIP steels (at 293 K unless otherwise noted). The Cantor alloy displays
fatigue behavior most similar to twinning-induced plasticity (TWIP) steels [61-65]. Austenitic stainless
steels with comparable microstructure show slightly lower threshold with similar Paris slopes of
approximately 3.5-4.5. The Al-containing HEAs demonstrated both higher Paris-regime slopes and fatigue
thresholds [35], which may be attributed largely to the as-cast microstructure in combination with the small
sample sizes that were used during testing.

40
Table 5.1: Comparison of fatigue crack growth testing parameters and results for data shown in Fig. 5.1
Alloy R Frequency Grain Size E Paris slope Threshold
[Hz] [µm] [GPa] (m) [MPa √m]
HEA – CrMnFeCoNi (293 K) 0.1 25 7 202 3.5 4.8
HEA – CrMnFeCoNi (198 K) 0.1 25 7 210 4.5 6.3
HEA – AlCrFeNi2Cu [35] 0.1 20 as-cast - 3.4 17
HEA – Al0.2CrFeNiTi0.2 [35] 0.1 20 as-cast - 4.9 16
XIP 1000-TWIP steel [61] 0.1 20 2 188 2.7 5.9
Fe-22wt%Mn-0.6wt%C
0.1 20 5 - 3.8 7.6
TWIP steel [62]
300-M steel (ISO250) [63] 0.05 50 20 205 2.5 3.6
304 steel [64] 0.1 20 - 200 3.8 15.8
304L steel [65] 0.1 35 40 193 5.8 4.8
316L steel [65] 0.1 35 38 193 4.9 3.3
(-) indicates information that was not reported in the literature

5.2. FUTURE WORK

The paucity of information in the literature on the fatigue behavior of HEAs and other
metallic systems provides ample opportunity for further experimental and theoretical
work to measure, predict, and understand the mechanistic variables at play. This
information in turn can be used not only to allow the use of a particular alloy system such
as the Cantor alloy with a high degree in confidence for performance expectations but to
also aid in the design of future alloy compositions.

With regards to work on understanding the fatigue behavior of the Cantor alloy,
several avenues of experimentation present themselves. One of the most promising is an
investigation into the fatigue behavior of the ‘medium entropy’ relative CrCoNi; this
alloy has been found to display similar or superior mechanical behavior, particularly
strength and fracture toughness, a similar temperature-dependence of said mechanical
properties, and a high propensity for deformation twinning. Another avenue of potential
future work is to expand the variety of testing conditions the Cantor alloy is subject to
including broader temperature range, both into the high temperature regime and
potentially even lower than 77 K (liquid He). Other test parameters could be varied as
well, such as a greater sampling of R-ratios to more fully characterize Schmidt and Paris
behavior of the Cantor alloy or larger sample size to allow testing into the higher, non-
linear region of the da/dN curve. Further, other environments could be simulated, such as
high salinity, to evaluate the performance of this material under harsher conditions in
which corrosion or oxidation may affect performance.

41
 Beyond compositional changes and testing conditions, other variants of this alloy
could be probed, with the most interesting potentially being varied grain size. The alloy
studied here displayed a rather fine and even microstructure, with average grain size <10
µm. The coarsening of the microstructure would likely further effect the fatigue behavior
of this system and may be a contributing factor to the large performance difference
between this alloy and that of others reported in the literature.

42
 Chapter 6
Conclusions
Just as its strength and toughness have been shown to improve with lowered
temperature, the fatigue behavior of the Cantor alloy was also found to display a certain
degree of temperature-dependent enhancement, revealing a greater resistance to crack-
propagation as temperature was lowered into the cryogenic regime. However, the
primary mechanism at work was found to differ from those responsible for the
improvement of its other low-temperature mechanical properties. While deformation
twinning has been theorized to play a role in the superior fracture toughness the Cantor
alloy displays at 77 K, little evidence for deformation twinning was observed in this
material. Deformation twinning was only observed at high ΔK and very low temperature
(77 K) and was only present in a small fraction of grains in proximity to the grain
boundary. As such, this mechanism does not likely play a role in the shifts of fatigue
threshold with temperature. Rather, strong evidence was found to indicate roughness-
induced crack closure serves as the dominant underlying mechanism of the observed
fatigue behavior shifts. This work additionally provides a solid foundation upon which
future studies may build, including but not limited to further mechanistic studies of the
role and activation of twinning under cyclic loading conditions, investigation of the
Cantor alloy’s fatigue performance under other environmental conditions, and
evaluation of the fatigue behavior of related alloys such as the ‘medium entropy’ CrCoNi.

In general, the Cantor alloy was found to demonstrate fatigue behavior comparable
or superior to other low temperature steel systems including austenic stainless steels and
TWIP steels under similar testing conditions. The dearth of fatigue data of HEAs prevents
much intra-family comparison, and the little data that is present in the literature describes
systems of radically different composition or microstructure. Despite the increasing
interest in the HEA family, fatigue studies of these materials remain underrepresented in
the literature. As such, this study provides an important step towards closing the gap
between scientific research and practical application through the characterization of the
fatigue crack-propagation behavior displayed by one of the most prominent and well-
researched members of the HEA family.

43
 References
[1] B. Cantor, I. T. H. Chang, P. Knight, A. J. B. Vincent, Microstructural development
in equiatomic multicomponent alloys, Mater. Sci. Eng. A 375-377 (2004) 213-218.
[2] J. W. Yeh et al, Nanostructured high-entropy alloys with multiple principle
elements: novel alloy design concepts and outcomes, Adv. Eng. Mater. 6 (2004)
299-303.
[3] J. W. Yeh, Alloy design strategies and future trends in high-entropy alloys, JOM -
J. Min. Met. Mat. S. 65 (12) (2013) 1759-1771.
[4] F. Otto, Y. Yang, H. Bei, E. P. George, Relative effects of enthalpy and entropy on
the phase stability of equatomic high-entropy alloys, Acta Mater. 61 (7) (2013)
2628-263.
[5] E. J. Pickering, N. G. Jones, High-entropy alloys: a critical assessment of their
founding principles and future prospects, Int. Mater. Rev. 61 (2016) 1743-2804.
[6] P. Chen et al, Fatigue behavior of high-entropy alloys: A review, Sci. China
Technol. Sci. 61 (2) (2018) 168-178.
[7] B. Gludovatz, E. P. George, R. O. Ritchie, Processing, microstructure and
mechanical properties of the CrMnFeCoNi high-entropy alloy, JOM - J. Min. Met.
Mat. S. 67 (2016) 2262-2270.
[8] B. Gludovatz et al, A fracture-resistant high-entropy alloy for cryogenic
applications, Science 345 (2014) 1153-1158.
[9] M. J. Yao, K. G. Pradeep, C. C. Tasan, D. Raabe, A novel, single phase, non-
equiatomic FeMnNiCoCr high-entropy alloy with exceptional phase stability and
tensile ductility, Scripta Mater. 72-73 (2014) 5-8.
[10] F. Otto et al, The influences of temperature and microstructure on the tensile
properties of CoCrFeMnNi high-entropy alloy, Acta Mater. 61 (2013) 5743-5755.
[11] A. Gali, E. P. George, Tensile properties of high-and medium-entropy alloys,
Intermetallics 39 (2013) 74-78.
[12] Y. F. Ye, Q. Wang, J. Lu, C. T. Liu, Y. Yang, High-entropy alloy: challenges and
prospects, Mater. Today 19(6) (2016) 349-362.
[13] Y. Zhang, X. Yang, P. K. Liaw, Alloy design and properties optimization of high-
entropy alloys, JOM - J. Min. Met. Mat. S. 64 (7) (2012) 830-838.
[14] Y. J. Zhou, Y. Zhang, Y. L. Wang, G. L. Chen, Solid solution alloys of
AlCoCrFeNiTix with excellent room-temperature mechanical properties, Appl.
Phys. Lett. 90 (18) (2007) 181904.
[15] S. T. Chen et al, Microstructure and properties of age-hardenable Al xCrFe1.5MnNi0.5
alloys, Mater. Sci. Eng. A 527 (21-22) (2010) 5818-5825.

44
[16] C. Y. Hsu, C. C. Juan, W. R. Wang, T. S. Sheu, J. W. Yeh, S. K. Chen, On the superior
hot hardness and softening resistance of AlCoCr xFeMo0.5Ni high-entropy
alloys, Mater. Sci. Eng. A 528(10-11) (2011) 3581-3588.
[17] O. N. Senkov, J. M. Scott, S. V. Senkova, D. B. Miracle, C. F. Woodward, (2011).
Microstructure and room temperature properties of a high-entropy TaNbHfZrTi
alloy, J. Alloys Compd. 509 (20) (2011) 6043-6048.
[18] C. J. Tong et al, Mechanical performance of the AlxCoCrCuFeNi high-entropy alloy
system with multiprincipal elements, Metall. Mater. Trans. A 36 (5) (2005) 1263-
1271.
[19] C. Y. Hsu, J. W. Yeh, S. K. Chen, T. T. Shun, Wear resistance and high-temperature
compression strength of Fcc CuCoNiCrAl0.5Fe alloy with boron addition, Metall.
Mater. Trans. A 35 (5) (2004) 1465-1469.
[20] M. H. Chuang, M. H. Tsai, W. R. Wang, S. J. Lin, J. W. Yeh, Microstructure and
wear behavior of AlxCo1.5CrFeNi1.5Tiy high-entropy alloys, Acta Mater. 59 (16)
(2011) 6308-6317.
[21] M. H. Chuang et al, Intrinsic surface hardening and precipitation kinetics of
Al0.3CrFe1.5MnNi0.5 multi-component alloy, J. Alloys Compd. 551 (2013) 12-18.
[22] M. H. Tsai et al, Thermal stability and performance of NbSiTaTiZr high-entropy
alloy barrier for copper metallization, J. Electrochem. Soc. 158 (11) (2011) H1161-
H1165.
[23] Y. Zhang et al, Microstructures and properties of high-entropy alloys, Prog. Mater.
Sci. 61 (2014) 1-93.
[24] R. V. Marrey, R. Burgermeister, R. B. Grishaber, R. O. Ritchie, Fatigue and life
prediction for cobalt-chromium stents: a fracture mechanics analysis,
Biomaterials 27 (9) (2006) 1988-2000.
[25] K. Alagarsamy et al, Mechanical Properties of High Entropy Alloy Al0.1CoCrFeNi
for Peripheral Vascular Stent Application, Cardiovasc. Eng Technol. 7 (4) (2016)
448-454.
[26] A. Runciman, D. Xu, A. R. Pelton, R. O. Ritchie, An equivalent strain/Coffin–
Manson approach to multiaxial fatigue and life prediction in superelastic Nitinol
medical devices, Biomaterials 32 (22) (2011) 4987-4993.
[27] J. N. Yang, S. Chen, Fatigue reliability of gas turbine engine components under
scheduled inspection maintenance, J. Aircraft 22 (1985) 415-422.
[28] B. A. Cowles, High cycle fatigue in aircraft gas turbines—an industry
perspective, Int. J. Fracture 80 (2-3) (1996) 147-163.
[29] Y. Zhang et al, Microstructures and properties of high-entropy alloys, Prog. Mater.
Sci. 61 (2014) 1-93.
[30] M. C. Gao, J. W. Yeh, P. K. Liaw, Y. Zhang, High-Entropy Alloys, Springer
International Publishing (2016).

45
[31] M. A. Hemphill et al, Fatigue behavior of Al0.5CoCrCuFeNi high entropy alloys,
Acta Mater. 60 (2012) 5723-5734.
[32] M. A. Hemphill, Fatigue behavior of high-entropy alloys, Master’s thesis (2012).
[33] Z. Tang, T. Yuan, C.-W. Tsai, J. W. Yeh, C. D. Lundin, P. K. Liaw, Fatigue behavior
of a wrought Al0.5CoCrCuFeNi two-phase high-entropy alloy, Acta Mater. 99
(2015) 247-258.
[34] K. Liu et al, (2019), Fatigue behavior of ultrafine grained triplex Al 0.3CoCrFeNi high
entropy alloy, Int. J. Fatigue 158 (2019) 116-120.
[35] M. Seifi et al, Fracture toughness and fatigue crack growth behavior of as-cast high-
entropy alloys, JOM - J. Min. Met. Mat. S. 67 (2015) 2288-2295.
[36] N. Stepanov et al, Effect of cryo-deformation on structure and properties of
CoCrFeNiMn high-entropy alloy, Intermetallics 55 (2015) 8-17.
[37] F. Otto et al, Decomposition of the single-phase high-entropy alloy CrMnFeCoNi
after prolonged anneals at intermediate temperatures, Acta Mater. 112 (2016) 40-
52.
[38] C. C. Tasan, Y. Deng, K. G. Pradeep, M. J. Yao, H. Springer, D. Raabe, Composition
dependence of phase stability, deformation mechanisms, and mechanical
properties of the CoCrFeMnNi high-entropy alloy system, JOM - J. Min. Met. Mat.
S. 66 (2014) 1993-2001.
[39] N. G. Jones, J. W. Aveson, A. Bhowmik, B. D. Conduit, H. J. Stone, On the entropic
stabilization of an Al0.5CrFeCoNiCu high entropy alloy, Intermetallics 54 (2014)
148-153.
[40] D. Ma, B. Grabowski, F. Körmann, J. Neugebauer, D. Raabe, Ab initio
thermodynamics of the CoCrFeMnNi high entropy alloy: importance of entropy
contributions beyond the configurational one, Acta Mater. 100 (2015) 90-97.
[41] J. W. Yeh, Physical metallurgy of high-entropy alloys, JOM - J. Min. Met. Mat. S.
67 (2015) 2254-2261.
[42] D. B. Miracle, J. D. Miller, O. N. Senkov, C. Woodward, M. D. Uchic, J. Tiley,
Exploration and development of high entropy alloys for structural applications,
Entropy 16 (2014) 494-525.
[43] B. Cantor, Multicomponent and high entropy alloys, Entropy 16 (2014) 4749-4768.
[44] R. O. Ritchie, The conflicts between strength and toughness, Nat. Mater. 11 (2011)
817-822.
[45] Y. Zang et al, Microstructures and properties of high-entropy alloys, Prog. Mater.
Sci. 61 (2014) 1-93.
[46] G. Laplanche, O. Horst, F. Otto, G. Eggeler, E. P. George, Microstructural evolution
of a CoCrFeMnNi high-entropy alloy after swaging and annealing, J. Alloys
Compd. 647 (2015) 548-57.

46
[47] ASTM Standard E1820, Standard test method for measurement of fracture
toughness, ASTM International, West Conshohocken, PA (2014). www.astm.org.
[48] C. J. Gilbert, J. M. McNaney, R. H. Dauskardt, R. O. Ritchie, Back-face strain
compliance and electrical-potential crack length calibrations for the disk-shaped
compact-tension DC(T) specimen, JTEVA 22 (1994) 117-120.
[49] B. Gludovatz, S. E. Naleway, R. O. Ritchie, J. J. Kruzic, Size-dependent fracture
toughness of bulk metallic glasses, Acta Mater. 70 (2014) 198-207.
[50] K. Heckel, R. Wagner, The tensile fatigue behavior of CT-specimens with small
notch root radius, Int. J. Fatigue 11 (1) (1975) 135-140.
[51] ASTM Standard E647, Standard test method for measurement of fatigue crack
growth rates, ASTM International, West Conshohocken, PA (2013).
www.astm.org.
[52] R. Pippan, K. Haberz, H. P. Stüwe, The plastic deformation of fracture surface
contacts in fatigue, Eng. Fract. Mech. 53(3) (1996) 441-448.
[53] G. Laplanche, P. Gaduad, O. Horst, F. Otto, G. Eggeler, E. P. George, Temperature
dependencies of the elastic moduli and thermal expansion coefficient of an
equiatomic, single phase CoCrFeMnNi high-entropy alloy, J. Alloy Compd. 623
(2015) 348-53.
[54] A. Haglund, M. Koehler, D. Catoor, E. P. George, V. Keppens, Polycrystalline
elastic moduli of a high-entropy alloy at cryogenic temperatures, Intermetallics 58
(2015) 62-64.
[55] R. O. Ritchie, W. Yu, Short crack effects in fatigue: a consequence of crack tip
shielding, Small Fatigue Cracks, R. O. Ritchie and J. Lankford, eds., TMS-AIME
(1986) 167-189.
[56] S. Suresh, R. O. Ritchie, A geometric model for fatigue crack closure induced by
fracture surface roughness, Metall. Trans. A 13A (1982) 1627-1631.
[57] R. O. Ritchie, Mechanisms of fatigue crack propagation in metals, ceramics and
composites: role of crack tip shielding, Mater. Sci. Engng. 103 (1998) 15-28.
[58] L. Lawson, E. Y. Chen, M. Meshii, Near-threshold fatigue: a review, Int. J. Fatigue
21 (1999) S15-S34.
[59] R. A. Schmdt, P. C. Paris, Theshold for fatigue crack propagation and the effects of
load ratio and frequency, ASTM STP 536 (1973) 79-94.
[60] K. T. Venkateswara Rao, W. Yu, R. O. Ritchie, Fatigue crack propagation in
aluminum-lithium alloy 2090: part I. long crack behavior, Metall. Trans. A, 19A
(3) (1988) 549-561.
[61] T. Niendorf. et al, Fatigue crack growth - microstructure relationships in a high-
manganese austenitic TWIP steel, Mater. Sci. Eng. A 527 (2010) 2412-2417.

47
[62] P. Ma et al, Fatigue crack growth behavior of a course- and a fine-grained high
manganese austenitic twin-induced plasticity steel, Mater. Sci. Eng. A 605 (2014)
160-166.
[63] R. O. Ritchie, Near-threshold fatigue crack propagation in ultra-high strength
steel: influence of load ratio and cyclic strength, J. Eng. Mater.-T. ASME 99 (1977)
195-204.
[64] L. W. Tsay, Y. C. Liu, M. C. Young, D. Y. Lin, Fatigue crack growth of AISI 304
stainless steel welds in air and hydrogen, Mater. Sci. Eng. A 374 (2004) 204-210.
[65] C. Sarrazin-Baudoux, J. Petit, C. Amzallag, Near-threshold fatigue crack
propagation in austenitic stainless steel, Paper presented at ECF14, Cracow, 2002.
[66] J. W. Christian, S. Mahajan, Deformation twinning, Prog. Mater. Sci. 39 (1995) 1-
157.
[67] A.-L. Gloanec, G. Hénaff, D. Bertheau, P. Belaygue, M. Grange, Fatigue crack
growth behaviour of a gamma-titanium-aluminide alloy prepared by casting and
powder metallurgy, Scr. Mater. 49(9) (2003) 825-830.

48