2018 Book FractographyAndFailureAnalysis PDF
2018 Book FractographyAndFailureAnalysis PDF
2018 Book FractographyAndFailureAnalysis PDF
Fractography
and Failure
Analysis
Structural Integrity
Volume 3
Series editors
José A. F. O. Correia, Faculty of Engineering, University of Porto, Porto, Portugal
Abílio M. P. De Jesus, Faculty of Engineering, University of Porto, Porto, Portugal
Advisory editors
Majid Reza Ayatollahi, Iran University of Science and Technology, Tehran, Iran
Filippo Berto, Norwegian University of Science and Technology, Trondheim, Norway
Hermes Carvalho, Federal University of Minas Gerais, Pampulha, Belo Horizonte,
MG, Brazil
Alfonso Fernández-Canteli, University of Oviedo, Gijón, Spain
Matthew Hebdon, Virginia Tech, Blacksburg, USA
Andrei Kotousov, University of Adelaide, Adelaide, Australia
Grzegorz Lesiuk, Wroclaw University of Science and Technology, Wroclaw, Poland
Yukitaka Murakami, Kyushu University, Fukuoka, Japan
Shun-Peng Zhu, University of Electronic Science and Technology of China,
Chengdu, Sichuan, China
The Structural Integrity book series is a high level academic and professional series
publishing research on all areas of Structural Integrity. It promotes and expedites
the dissemination of new research results and tutorial views in the structural
integrity field.
The Series publishes research monographs, professional books, handbooks,
edited volumes and textbooks with worldwide distribution to engineers,
researchers, educators, professionals and libraries.
Topics of interested include but are not limited to:
– Structural integrity
– Structural durability
– Degradation and conservation of materials and structures
– Dynamic and seismic structural analysis
– Fatigue and fracture of materials and structures
– Risk analysis and safety of materials and structural mechanics
– Fracture Mechanics
– Damage mechanics
– Analytical and numerical simulation of materials and structures
– Computational mechanics
– Structural design methodology
– Experimental methods applied to structural integrity
– Multiaxial fatigue and complex loading effects of materials and structures
– Fatigue corrosion analysis
– Scale effects in the fatigue analysis of materials and structures
– Fatigue structural integrity
– Structural integrity in railway and highway systems
– Sustainable structural design
– Structural loads characterization
– Structural health monitoring
– Adhesives connections integrity
– Rock and soil structural integrity.
123
Jorge Luis González-Velázquez
Department of Engineering in Metallurgy
and Materials
Instituto Politécnico Nacional
Mexico City
Mexico
This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The study of fractured surfaces has been a fundamental part of materials research,
practically since the formalization of the study of material mechanical behavior in
the second half of the nineteenth century. Nevertheless, it was considered a minor
specialty, usually subordinated to fracture mechanics until the post-Second World
War years, when high impact fractures involving mechanical and structural com-
ponents, particularly in the aerospace industry, showed the usefulness of studying
fractured surfaces. Since the introduction of the term “Fractography” by Carl A.
Zapee at the 26th Annual Convention of ASM in 1944, it was recognized as a key
discipline for scientific research and new material development. But it was not until
the publication of the Metals Handbook, 8th Edition, Vol. 9 “Fractography and
Atlas of Fractography” in 1974, by the American Society for Metals International
(ASM), the leading organization in the publication of technical texts in the field of
metallic materials for engineering use, that Fractography became a consolidated
engineering specialty; likewise, the close relationship between fracture examination
and failure analysis encouraged ASM to publish, in 1987, the Metals Handbook,
Vol. 12, “Fractography,” the Metals Handbook, Vol. 11 “Failure Analysis and
Prevention,” and the Metals Handbook, Vol. 19 “Fatigue and Fracture,” completing
a series of handbooks related to fracture and material failure.
The fracture and failure analysis Metals Handbooks of ASM are comprehensive
compendiums of technical articles and numerous case studies, but their technical
level is often very high, and the amount of information contained in them is difficult
to process for those without previous experience in the area. This is the reason why,
nowadays, the industrial and academic areas need a textbook that may be used for
technical staff and students in their training for incursion into the fields of fracture
studies and failure analysis. It is precisely this need that motivated me to write the
book “Fractography and Failure Analysis,” with the intention of offering
easy-to-read material, even for people with a basic knowledge of engineering, and
to introduce the reader to the correct fracture examination methodology and to the
performance of effective failure analysis of components whose failure is associated
with fracture or cracking processes.
v
vi Preface
This book is the result of over 30 years of my own experience researching the
mechanical behavior of metals, fracture mechanics, and structural integrity, as well
as the performance of over 500 failure analyses of mechanical and structural
components. The original text comes from the Fractography Notes that I wrote for
my courses of Fractography and Failure Analysis that I have taught since 1992 at
the graduate school of the Metallurgy and Materials Department of the Instituto
Politecnico Nacional (National Polytechnic Institute) in Mexico City. For this
reason, this book cites very few references, since all of the photographic and
theoretical material is original. This book also completes my trilogy of books on
mechanical behavior and fracture, the first two being “Mechanical Metallurgy” and
“Fracture Mechanics,” both published in Spanish by Editorial Limusa, Mexico.
Chapter 1 begins with a description of what, to my judgment, is the correct
methodology for the examination of fractured physical components. This topic is
introduced because of the nonexistence of a complete description of ASTM E2332
recommended practice for the examination of fractured components, so as to train
young engineers, failure analysis apprentices, and novice fracture researchers, who
usually make severe mistakes and errors in the interpretation of fractures, or worse,
irreversibly alter the examined pieces, causing serious problems, especially when
the examined item is part of a judicial investigation. For these reasons, this chapter
describes the best practices for the examination and handling of fractures, including
a practical guide for quality photography, cutting, cleaning, and preservation of
fractures. The first chapter also includes a description of the making of fracture
replicas, since these activities are an essential part of the fractographic examination.
Chapter 2 presents the basic concepts for the scientific study of fractured sur-
faces, their classification, and a description of the main fracture lines and roughness
that are visible to the naked eye. With this information, the reader will be able to
determine the propagation sequence of a fracture, as well as identify the fracture
mechanism and the effect of the stress state and mechanical properties on the
fracture characteristics.
Chapters 3, 4, and 5 focus on the study of brittle and ductile, fatigue and
environmentally assisted fractures, starting with a description of the general char-
acteristics of the fracture mechanisms and continuing with a description of fracture
features of each mode of fracture at both the macroscopic and microscopic levels.
The text is written in such a way that the reader will be able to learn how to identify
the fractured surfaces and determine the factors that influence the fracture behavior.
Occasionally, the topic sequence may vary from one chapter to another, in order to
give clarity and logical sequence to the explanations contained herein.
Chapter 6 is devoted exclusively to the description of a failure analysis proce-
dure that I developed based on my own experience, but that is in accordance with
the procedure described in ASTM E 2332-04 “Investigation and Analysis of
Physical Component Failures,” nowadays the only internationally recognized
practice for failure analysis of mechanical and structural components. This stan-
dard, however, only presents a list of the activities necessary to carry out a failure
analysis, giving no explanation or description for the assorted steps (ASTM E
2332-04 standard consists of two pages). Therefore, in this chapter, the objectives
Preface vii
ix
x Contents
3.1.3 Pseudo-Cleavage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.1.4 Cleavage in Amorphous Materials . . . . . . . . . . . . . . . . . . 57
3.1.5 Intergranular Brittle Fractures . . . . . . . . . . . . . . . . . . . . . . 58
3.2 Ductile Fractures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.1 Macroscopic Appearance of Ductile Fractures . . . . . . . . . . 60
3.2.2 Ductile Fracture Mechanism . . . . . . . . . . . . . . . . . . . . . . . 63
3.2.3 Void Nucleation and Growth Mechanisms . . . . . . . . . . . . 66
3.2.4 Ductile-Brittle Transition . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2.5 Ductile-Brittle Combined Fractures . . . . . . . . . . . . . . . . . . 69
4 Fatigue Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.1 General Aspects of Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.2 Fractography of Stage I Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.2.1 Macroscopic Characteristics of Stage I Fatigue . . . . . . . . . 78
4.2.2 Microscopic Characteristics of Stage I Fatigue . . . . . . . . . . 79
4.2.3 Fatigue Crack Nucleation Mechanisms . . . . . . . . . . . . . . . 79
4.3 Fractography of Stage II Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.3.1 Macroscopic Characteristics of Stage II Fatigue . . . . . . . . . 84
4.3.2 Microscopic Characteristics of Stage II Fatigue . . . . . . . . . 86
4.3.3 Striation Formation Mechanisms . . . . . . . . . . . . . . . . . . . . 90
4.3.4 Fractographic Characteristics of Stage II Fatigue
Fracture in the Absence of Striations . . . . . . . . ........ 92
4.4 Corrosion-Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ 93
5 Environmentally-Assisted Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . 97
5.1 General Aspects of Environmentally-Assisted Fracture . . . . . . . . . 97
5.2 Fundamentals of Metal Corrosion . . . . . . . . . . . . . . . . . . . . . . . . 100
5.3 Stress Corrosion Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
5.3.1 Mechanical Aspects of Stress Corrosion Cracking . . . . . . . 104
5.3.2 Characteristics of SCC Fractures . . . . . . . . . . . . . . . . . . . . 106
5.3.3 SCC Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.4 Creep Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.4.1 General Aspects of Creep Fracture . . . . . . . . . . . . . . . . . . 111
5.4.2 Creep Fracture Mechanism . . . . . . . . . . . . . . . . . . . . . . . . 113
5.4.3 Creep Crack Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.5 Hydrogen-Induced Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.5.1 Hydrogen-Induced Cracking Mechanism . . . . . . . . . . . . . . 117
5.5.2 Fractography of Hydrogen-Induced Cracking . . . . . . . . . . 118
5.5.3 Mechanical Aspects and Kinetics of Hydrogen-Induced
Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6 Failure Analysis of Fractured Components . . . . . . . . . . . . . . . . . . . . 125
6.1 Definitions of Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.2 Procedure for the Analysis of Mechanical Failures . . . . . . . . . . . . 127
Contents xi
Abstract This chapter describes the adequate methodology for the examination of
fractured components, with emphasis being given to the specific features of frac-
tured components that have to be recorded for further analysis. The correct pro-
cedures for the handling, cutting, cleaning and preservation of fractures are
described in detail. Recommendations for taking good photographs of fractured
surfaces are given, including a description of the fundamentals of digital photog-
raphy. The chapter ends with a description of the elaboration of fracture replicas.
1.1 Introduction
Fractography is the study of fractured surfaces for the purpose of learning about
their topographic characteristics, identifying the types of fracture, understanding
what causes fracture, and eventually determining their mechanisms. Nowadays, it is
a very useful tool in materials science and engineering, as well as in failure analysis.
In materials science and engineering, by knowing the fracture mechanisms, we can
identify the characteristics of a given material that increase its fracture resistance, so
stronger materials can be designed, while in failure analysis, we can make the
essential determination of the cause of the failure.
For the study of fractography, it is necessary to have a technical definition of
what a fracture is so as to distinguish it from other surfaces. The best definition is:
fracture is the process of separation or fragmentation of a solid body under the
action of loads or stresses, thus creating new surfaces, which are referred to as the
“fractured surface”. Based on this definition, it is clear that the term “fracture” will
always involve at least two new surfaces, which are the objects of study within
fractography. Furthermore, a fracture is the result of the action of loads or stresses,
thus it is part of the field of mechanical behavior. There are several ways to separate
a solid material into pieces to create new surfaces, like flame cutting, erosion,
chemical dissolution, etc., but they do not involve the direct action of stresses, and
therefore they are not fractures.
View of a crack at the macroscopic scale in a free View of a crack at the microscopic scale,
surface. Note that the crack is viewed as a line. where it can be seen that it is formed by two
surfaces meeting at a tip.
that they are the intersection of the fracture planes with the free surface. The greater
the crack opening, the thicker the line will appear. The figures below help to
describe this (Fig. 1.1).
For its study, Fractography is supported by several procedures and techniques
that can be summarized as follows:
(a) Visual examination. Carried out just by looking with the naked eye or with the
aid of a magnifying glass, or low power magnifying lenses. The aim of the
visual examination is to determine the overall appearance of the fractured
surface, identify the fracture lines, identify the zones of different fractographic
appearance and identify aspects such as associated plastic deformation, number,
location and arrangement of cracks, corrosion, coatings, stress concentrators
(i.e., grooves, pits, cavities, sharp ends and holes), as well as dents, gouges,
physical damage, wear and other features of interest.
(b) Stereoscopic viewing. Basically the same scope as visual inspection, but with
greater detail, since the typical stereoscopes allow magnifications from
10 to 70.
(c) Microscopic examination. Generally done through Scanning Electron
Microscopy (SEM). This is the most powerful tool for examining fractured
surfaces, since it allows for the observation of the microscopic features, as well
as compositional details that help to reveal the fracture mechanism and the
interactions between the fracture mechanism and the microstructure and the
environment. It also shows fundamental features such as the microscopic crack
path, micro-roughness and micro-deformation. SEM typically provides mag-
nifications up to 100,000, although the practical limit for metals is of about
10,000 due to a phenomenon called “hollow resolution.” In addition to its
great magnification power, SEM possesses other advantages, such as: great
depth of field, several signals that, by themselves or combined, provide images
that reveal compositional, electrical conductivity, magnetic and other analytical
features.
4 1 The Fractographic Examination
(d) Transmission Electron Microscopy (TEM). Originally used for the observation
of replicas, nowadays TEM is rarely used for fractography, for which its main
use was to study the fracture underlaying deformation micro-mechanisms (by
observing the dislocation arrays) on thin layers observed under ultra-high
voltage devices (more than 1000 kV).
(e) Scanning Transmission Electron Microscopy (STEM). In this technique, the
image is formed by the combination of a scanning and a transmitted beam. This
technique has a scope similar to that of TEM, but with the advantage that it
allows for the fracture roughness with the underlying microstructure to be
related.
(f) Surface Microanalysis. This includes energy dispersive (EDS) and wave length
(WDS) X-ray analysis, Auger spectrometry and other techniques for thin layer
and particle chemical analysis of second phases and corrosion products on the
fractured surfaces.
It must be remembered that a fractured surface is delicate and unique, and
therefore the examination should be carried out with great care. If possible, without
cutting, damaging or altering the fractured surface. A permanent record of the
fractured surfaces must be kept, through photographs or replicas, and proper
preservation of the fractured pieces must be done for further study.
In the next sections, topics related to macroscopic fracture examination will be
described, as well as recommendations for their cutting, preservation and care,
along with some practical aspects of taking photographs.
Strictly speaking, each fracture is unique and unrepeatable, so even when two
identical pieces are fractured under similar conditions, the random variations of the
microstructure and mechanical properties, as well as the interactions of the crack
with the changing environment, will cause a fracture that can never be identical to
any other one. The aforementioned condition means that any fractographic report
should contain an exact and detailed description of the fractured piece, with all of
the photographs that indicate the features of interest. To accomplish this, the exam
must be done following a previously outlined plan and performed with strict
attachment to the appropriate methodology.
It is important to point out that a fractured surface examination can be performed
on a piece that became fractured either in service, during its manufacture or in a
laboratory under controlled conditions. In any of these cases, the fractographic
examination will become part of an investigation, with the objective of determining
the fracture mechanism or, likewise, the effect of specific factors on the fracture
mechanism or the fractographic features.
Whichever the case, failure analysis or research, the objectives of the fracto-
graphic examination are:
1.2 Methodology of the Fractographic Examination 5
1
Roughness can be measured as the average height difference between crests and valleys on the
fractured surface.
6 1 The Fractographic Examination
Cutting for
Preservation of fracture
surface and all fragments Microscopic Exam
and Test Samples
(Only if necessary)
One of the most important aspects of the fractographic examination sequence is the
handling and care of fractures, in order not to damage or alter the delicate char-
acteristics of the fractured surface. The following guidelines are recommended for
the correct handling and care of a fractured surface:
1. Do not try to put two fractured surfaces together again.
2. Do not mark the fractured surface.
3. Do not remove any fragments from the surface.
4. Do not apply any hard instruments or sharp edges to the fractured surface.
5. Do not touch the fractured surface.
6. Do not use acids or any corrosive substance for cleaning.
7. Do not use cloth, fiber or any sort of paper that may leave lint, residues or
filaments.
8. Do not apply masking tape, glue, adhesives or paints to the fractured surfaces.
These restrictions are meant for the conservation of a fracture for microscopic
examination, in which fine details are key to succeeding in the correct identification
of the fracture mechanism. Carelessness or negligence may permanently alter such
fine details.
The photographs in Fig. 1.3 shows some examples of mishandling of fractured
surfaces.
8 1 The Fractographic Examination
Fig. 1.3 a Original fracture. b Same fracture exposed to air for several days. c Same fracture with
mechanical damage. d Corroded fracture surface
As the reader may have noticed, cutting out a fractured piece is quite often required,
because the piece may be too big to be easily handled, or simply because it will not
fit into the SEM specimen chamber. However, cutting should always be the last
resort, and whenever possible, the fractured surface must be kept whole. If the
fractography work is part of a failure analysis, it is advisable that the analyst travel
to the site where the fracture is located to perform the macroscopic examination,
take pictures, and indicate the required cut-offs for the transfer of the piece to the
lab. Handling and preservation must be specified for the people in charge of
delivery of the specimen.
The most common cutting methods are:
• Blow torch- This is the best option for large metallic pieces, the most common
method being the oxyacetylene blow torch. In this case, the extension of the
heat-affected zone must be taken into consideration so that the zone of interest is
not heat-damaged. Whatever the case, the heat-affected zone should be removed
1.3 Cleaning and Care of Fractures 9
Fig. 1.4 Metal drops sputtered over a fractured surface during flame cutting
• Water jet- This method uses a high-pressure water jet (about 70,000 psi).
Relatively soft materials are cut with plain water, but a dosage of abrasive may
be added for quicker cutting of very hard materials. Water jets can cut almost
any material, including ceramics and glass. The cutting is fast, for example, steel
slabs of more than 10 cm thickness can be cut at speeds of over 10 cm/min. This
technique produces neither heat nor distortion, and the system can be mounted
on a computer-controlled table for cutting complex forms.
• Electro-Erosion- This is a cut made by means of an electric arch, which jumps
out from a high voltage copper electrode onto a piece that is ground-connected.
The piece is immersed in a dielectric oil, in order to give stability to the arch.
The cutting is done according to the shape of the electrode. This method enables
very precise cuts and does not damage the material, but it is only applicable to
conducting metals.
The next practical aspect of top importance within the macroscopic examination is
cleaning. Cleaning is necessary to remove dirt or corrosive products attached to the
fractured surface that obstruct the observation of fine surface details. If observation
of the residue on the surface is not wished or if it has already been analyzed, the
fractured surface can be cleaned following the sequence given below, interrupting it
at the moment that an adequate surface cleanness is achieved.
In addition to using cleaning substances, it is advisable to make use of an
ultrasonic bath, which is a tub made of stainless steel, connected to an ultrasonic
vibrator located on the outer side of the bottom of the tank, as shown in the figure
below. In most commercial equipment, the vibration rate is fixed and the time is
pre-set in a timer that shuts it off automatically. The size varies depending on the
size of the piece to be cleaned and the more sophisticated models have pro-
grammable cycles of vibration time and heating.
When cleaning metallic pieces, any direct contact of the piece with the tank must
be avoided so as not to create a galvanic pair. To do this, the tank is partially filled
with water and the piece is placed inside a glass beaker filled with the cleaning
solution, as shown in Fig. 1.5.
The recommended sequence for cleaning fractured surfaces is:
1. Blow off- This must be done with dry air at room temperature. Compressed air
bottles can be used, such as those used for cleaning electronic equipment,
however, condensation of humidity must be prevented, as canned air is too cold.
2. Brushing- Preferably done with a paint brush of organic bristles.
3. Cleaning by immersion or soaking- Organic solvents like Xilene, Naphtha,
Toluene, acetone, alcohol or gasoline may be used. Solvents should preferably
be volatile or soluble with other solvents so they can be easily removed and
leave no residues when evaporated. Ethylic alcohol may be used as a first
1.3 Cleaning and Care of Fractures 11
BEAKER
TUB CLEANING
SOLUTION
WATER
PIECE
Timer Power
source PIEZOELECTRIC
VIBRATOR
Preserving a fractured surface, provided that the material is stable enough, will only
require cleaning and storage in a cool, dry place, such as on a shelf or in a box.
Storage in dryer bowls is not generally recommended, unless there is a
humidity-absorbent material present, since dry air is very corrosive for most metals.
For slightly air reactive metals, such as steel or copper alloys, the oil known as
“WD-40” is an excellent option. The fractured surface must be thoroughly sprayed,
letting the excess drain away. Very reactive materials or pieces that will be stored
for long times may be coated with a lacquer or another inert substance. They have
to be waterproof and easily removable with solvents chemically compatible with
the fractured piece. Transparent nail enamel is a good option for most metallic
alloys.
MM ¼ ER=IR;
where ER is the eye resolution and IR is the instrument resolution. For a healthy
human eye, ER is about 0.1 mm, thus, if the resolution of an instrument (camera
lens or microscope) is, for example, 0.001 mm, the maximum magnification to be
obtained would be 1000. A greater magnification will make objects separated by
less than 0.0001 mm look as if they were one object or a blurred point; therefore, it
is useless to try to obtain images with greater magnification than the one provided
1.4 Photographing Fractures 13
Fig. 1.6 Example of hollow resolution. a Photograph at the original magnification, as taken in an
SEM. b Enlargement of the original image. Note that the enlargement makes the image bigger, but
does not reveal fine details
2
Pixel is an acronym for ‘picture element,’ which is the dot or square of a given color and
brightness that forms an image.
14 1 The Fractographic Examination
the sensitivity, as it allows more light to enter) and the intrinsic sensitivity of the
image sensor that makes up the digital recording system. In photographic films, the
ASA/ISO number determines the light sensitivity, resolution (grain size) and
contrast. A low ASA/ISO means lower sensitivity, but higher resolution, and vice
versa. In digital photography, light sensitivity or ASA may be adjusted through the
camera settings, but the limits are set by the manufacturer (a minimum/maximum
ASA). In addition, the digital images can be improved in regard to brightness and
contrast by image processing systems.
Magnification- Magnification is the size at which an image is originally cap-
tured, but not the print size. Magnification is determined by the optic system, and
the purpose in fracture photography is to get the greatest magnification without
losing resolution (avoiding hollow magnification). Generally speaking, the greater
the magnification needed, the more complex and costly the system will be and the
more care will be necessary in taking the photo, especially considering vibrations,
stability and cleanliness of the optic system. Due to the hollow magnification effect,
the best option will be to get the picture at the desired magnification from the very
beginning.
Depth of field- This is the distance along the optical axis at which objects remain
in focus. Depth of field depends on the optic system, and the rule is that “the greater
the magnification, the smaller the depth of field.” In any optic system, the depth of
field is controlled by the diaphragmatic opening of the objective lens; the smaller
the opening, the greater the depth of field. In fractography, it is advisable to get
greater depths of field, but this always leads to a compromise with magnification, so
the fractographer will have to decide what magnifications are best in order to get the
desired depth of field.
Illumination- Illumination determines the brightness and the contrast, which,
combined with focus and resolution, make up the picture’s quality. For fractogra-
phy, first of all, an adequate source of light must be chosen. Of course, the best
source is natural light (sunlight), as it contains the whole color spectrum visible to
the human eye, but it has the limitation that it cannot be controlled, therefore it is
rarely used for professional photography, which prefers artificial lighting. The most
common artificial sources of light are: incandescent tungsten lights, halogen lamps,
fluorescent lights, LED and xenon lights. All of these have the advantage that the
intensity and direction can be controlled. Nonetheless they also have their disad-
vantages: first, as they come from a relatively small source, the illumination is not
even, and second, they provide a limited range of wavelengths, which affects the
contrast, especially in color photography. Both tungsten and xenon lights produce
yellowish tonalities, so the details in this color will fade, whereas fluorescent, LED
and xenon lights give off an excess of blue light or ultraviolet light, thus producing
a whitish tonality over color, all of which gives them a “washed out” aspect. In
practice, such problems may be corrected or minimized by filters, diffusers or other
accessories. For the fractographic job, the best results are obtained by trial and
error, using different light sources. The advantage of digital photography is that the
contrast and brightness can be further adjusted or corrected by image processing
software.
1.4 Photographing Fractures 15
Fig. 1.7 Effect of the illumination angle on contrast in fracture photography. a Vertical
illumination. b Oblique illumination from the left side. c Oblique illumination from the right side
16 1 The Fractographic Examination
1.5 Replicas
Finished replica
Replicant material
Piece
Fig. 1.9 Schematic representation of the process for obtaining a one-step replica
Step 1
Replicant material Direct replica
PIECE
REPLICA 1
Fig. 1.10 Schematic representation of the process for obtaining a two-step replica
18 1 The Fractographic Examination
(a)
(b) (c)
Fig. 1.11 Comparison of replicas: a Original fracture. b One step replica made of a light polymer.
c Two step replica made of plaster
allowed to harden for a few minutes. This material is very stable, so it can be
cut, painted or coated. For its observation in the SEM, it will be necessary to
metalize it in order to make it into a conductor, and it should be observed at low
voltage (5–10 kV). Its resolution is good for mid-magnifications (up to 1000).
• Alginate- Also a material for dental use, this is cheap and easy to use. Its
resolution is not as good as silicon, but it is good enough to carry out macro-
scopic examinations. It comes in the form of a powder and is mixed with water
until a viscous fluid is obtained. The fluid is poured onto the surface to be
reproduced without applying any pressure or heat. Once hardened, it must be
carefully detached, because it is not mechanically resistant. To improve its
appearance, it might be metalized or painted. It is good for two-step replicas.
• Plaster- For two-step replicas and big pieces, plaster may be a good option,
though its resolution is very limited. It is better to use high quality plasters for
dental moldings instead of construction plaster.
• Wax- This is not an engineering material, but it can be used when there is no
other option available. The wax is heated until it becomes fluid and then is
poured onto the surface to be reproduced. It hardens after cooling off, so it can
be removed carefully. The replica must be handled with extreme care, because
wax is a weak material and its boiling point is close to room temperature,
meaning that it may melt, even while being held with bare hands.
The appearance of all replicas is improved substantially when they are metalized
or painted. Metalizing through evaporation of a noble metal (gold, platinum or
rhodium) or graphite is the best option. The evaporation technique most recom-
mended is by incandescence in vacuum bells, because the heat generated with other
evaporation methods, such as plasma, may damage the replica. The following table
shows the most common defects in replicas, their causes and corrective actions
(Table 1.1).
Table 1.1 The most common defects on replicas, causes and corrections
Defect Cause Correction
Filaments Adherence of replicating material due Wait until the replica fully hardens
and to removal before full hardening
tongues
Porosity Trapped air Avoid excessive stirring or replica
material that is too soft
Poor Application of the replica material Follow the manufacturer’s
resolution after it has begun to harden recommendations for replica material
preparation
Voids Improper application. Application of Apply the replica material in a fluid
the replica material after it has begun and uniform fashion
to harden
Tearing Replica material trapped in cavities Make the replica from areas without
without an exit angle cavities. If the tearing is excessive,
the surface is not adequate for
replication
Fractures Overly brisk replica extraction or Remove the replica carefully and
removal of the replica when the after it fully hardens
material is not fully hardened
Chapter 2
Elements of Fractography
% AREA REDUCTION
BRITTLE
FRACTURE DUCTILE FRACTURE
Little or no plastic
deformation
Neck formation
Generalized 100%
The plastic R. A.
fragments deformation
adjust
perfectly
σ σ σ σu
σF = σo σF < σo
σO σO
σF = σu
ε ε ε
10-1 10-1
Ductile
Ductile
σ/E Rupture
ε s-1
High
temperature
fracture
Creep intergranular
fracture
Cleavage
10-6 10-6
0 1.0 0 1.0
T/Tf T/Tf
Fig. 2.2 Ashby maps of fracture mechanisms (schematic representation, not showing actual data)
24 2 Elements of Fractography
cleavage, whereas brittle forms of fracture are favored by low temperatures and high
strain rates. High temperature fracture (cavitation and viscous fluid) is favored by
low strain rates.
τ Mohr‘s Pure
circle in
torsion
τ max pure shear
Direction of
σ max
Fracture
σ max
plane
Fig. 2.3 Schematic representation of a fracture in pure shear of a brittle material. Note that the
fracture plane is oriented 45° from the twisting direction, which is the direction of the maximum
principal stress
Pure shear
τ
Mohr’s
circle
Pure
τmax torsion Direction of
τ max
σ max
Fracture
plane
Fig. 2.4 Schematic representation of a fracture in pure shear of a ductile material. Note that
fracture plane is parallel to the twisting direction, which is in the direction of the maximum shear
stress
is in the transversal plane. Thus, the fracture plane will be at 45° from the bar´s
longitudinal axis, as shown in Fig. 2.3. The opposite behavior is observed if a play
dough clay bar is twisted up to its rupture. Play dough is a ductile material, thus it
will fail by shear. In a pure shear stress state, the maximum shear stress is parallel to
the direction of twisting, therefore the fracture will be in a plane perpendicular to
the bar´s longitudinal axis, as shown in Fig. 2.4.
Experiment 2: Break a play dough bar in pure tension, with the load parallel to
the longitudinal axis of the bar. In this case, the material will start flowing plasti-
cally in the direction of the maximum shear stress, which will be at 45° from the
tension axis. Assuming that the material will behave as an ideal plastic (no strain
hardening) and that it is homogeneous, the transversal section will be continuously
reduced until reaching practically 100% of area reduction, whereupon the piece will
26 2 Elements of Fractography
σ
DIRECTION
σmax OF τmax
Fig. 2.5 Schematic representation of the fracture in pure tension of a ductile material. Note that
plastic flow occurs in the direction of the maximum shear stress
split into two parts. The sides of the final rupture zone will show a tilt close to 45°,
as shown in Fig. 2.5.
Fracture mechanics. The local character of the fracture prevents addressing
fracture stress as if it was a material property, and that is why Fracture Mechanics
was developed. The main achievement of Fracture Mechanics is that the conditions
of initiation and propagation of a crack can be established by a single parameter that
involves the loads or stresses acting on the body, the size and orientation of the
crack, and the geometric features of the component. Knowing the parameter that
controls the fracture helps us to understand the mechanical conditions that led to a
fracture and the way in which it happened.
The parameter that mechanically characterizes a fracture is set by the extension
of the plastic zone prior to failure, according to the following categories:
(a) Linear Elastic Fracture Mechanics- The plastic zone is small and surrounded
by elastically deformed material. Here, the parameter controlling the fracture is
the stress intensity factor (K).
(b) Elasto-Plastic Fracture Mechanics- If the plastic zone extends along the
ligament, but limits itself to a relatively narrow band, it is then said that the
fracture is elasto-plastic. The J Integral or the Crack Tip Opening Displacement
(CTOD) characterize the crack’s behavior.
(c) Plastic Collapse- If the plastic zone extends up to most of the body´s volume,
the fracture is then controlled by the stress in the ligament, called effective
stress.
Generally speaking, depending on the type of fracture and the size of the plastic
zone, when one of these parameters reaches a critical value, the instability begins,
that is, the sudden and catastrophic propagation of the crack. The critical fracture
parameter is usually a material´s property, and it depends on its chemical compo-
sition, the microstructure and the environment. The resistance to fracture is
then defined as fracture toughness. Table 2.1 shows the characteristic fracture
2.2 Mechanical Aspects of Fracture 27
The description of the fracture phenomenon introduced so far allows for the pro-
posal of a general model of fracture. Such a model views a solid body with a
reduced transversal area and subject to a tension load perpendicular to the plane of
the reduced transversal area. The transversal area reduction is represented by a
geometric discontinuity, which creates a stress concentration zone, thereby modi-
fying the material´s local behavior. In this zone, the first fracture stage takes place,
known as the Stage I of crack nucleation. Once the crack nucleus has been
formed, the Stage II of crack propagation occurs. It is during Stage II that the
main fractographic characteristics will be produced.
Right in front of the growing crack, there is the process zone, which is the region
at the crack tip where the fracture mechanisms takes place. The fracture mechanism
entails all plastic deformation processes and atomic bond ruptures that lead to the
creation of a fractured surface. In the fracturing process, there are two important
zones: the fractured surface and the process zone. Depending on what goes on in
the process zone, it will give the topographic characteristics of the fractured surface.
In other words, the fractured surface is what results from what happens in the
process zone. In the process zone, the plastic deformation is intense and the state of
the stress is usually more complex than the state of stress in the rest of the piece.
A strong interaction with the environment will take place, at both the process zone
and the fractured surface, due to the high reactivity with the deformed and fracture
28 2 Elements of Fractography
material. The last zone of the fracture is called Stage III of final failure, and it is
the zone where the final separation occurs. Because of the high stress concentration,
high strain rate and the major changes in geometry when the fracture is close to final
separation, the state of stress and strain change significantly in relation to the state
of the process zone in Stage II and, therefore, the fracture surface in Stage III will
show abrupt changes of orientation, roughness, secondary cracks, and so forth.
Figure 2.6 shows the general model of fracture described above. Understanding
of this model and correct identification of the fractographic characteristics will
enable the identification of each fracture stage and will help in determining the
sequence of the fracture, this in addition to providing the basis for a more detailed
study. The study of each fracture stage has different practical benefits, as they all
feature particular characteristics. Table 2.2 summarizes the study aims for each
fracture stage.
Load or stress
Stress concentrator
Crack
Initiation
zone Process
zone
Crack front
Fracture surface
Fig. 2.6 General fracture model
2.4 Fracture Lines 29
Fracture lines are actually changes in the fractured surface’s roughness that look
like lines because of the shadow formed under oblique light conditions. In other
cases, they can be the borderline between two regions with different surface
appearance, color, texture or roughness. Such changes may be caused by a different
fracture mechanism, environmental changes or microstructural changes. The main
lines of fracture are shown in Fig. 2.7.
2.4.2 Chevrons
These are V-shaped ridges appearing on the fractured surface of ductile and near
brittle materials. The vortex of the V points towards the crack initiation location, so
the propagation direction is the opposite. An example of chevrons is shown in
Fig. 2.8.
Chevrons form due to a curved crack propagation front; the more curved the
crack front, the more visible the chevron. The formation of a curved crack prop-
agation front is associated with the combination of plane stress with a plane strain
condition in the transversal section of the fractured piece. In a mid-thickness plate,
there is always a combination of plane stress with plane strain. The plane stress
zone is found on the lateral surfaces of the plate and the crack growth rate is slower
than in the plane strain zone located at the center of the plate. This is because the
toughness in the plain strain is less than that in the plane stress, due to the plastic
constriction phenomenon. A practical estimation shows that fracture toughness in
plane stress is 2.5 times the value of that in plane strain. The variation of fracture
toughness across the thickness in a mid-thickness plate is shown in Fig. 2.9.
30 2 Elements of Fractography
Ridges
Ratchet mark
Beach mark or
Clam Shell mark Shear lip
Chevron
Propagation Initiation
Fig. 2.8 Chevrons in the fracture of a mid-thickness, high strength steel plate
2.4 Fracture Lines 31
Figure 2.10 shows the hypothetical profile of the plastic zone through thickness
when there is a combination of plane stress and plane strain. The central zone
(which is in plane strain) is the first to fracture, due to its lower toughness. When
the fracture occurs in the middle thickness region, a small step is formed because of
the out-of-plane deformation and the crack front advances to the lateral surfaces of
the plate as the crack moves forward, initiating a step that will form the chevron.
This is a cyclical process, in the sense that while the crack front reaches the lateral
sides, the fracture is taking place a little ahead of the formed chevron, thus forming
a new chevron.
It has been found that the best combination of plane stress and plane strain is
obtained when the thickness (B) approaches the value given by
2
KIC
B ;
r0
Chevron: Ridge
formed by the
propagation from the
Plastic zone Crack front center to the sides
Fracture
Plane strain surface
Plane stress
Free surface
Fig. 2.10 Chevron formation due to the combination of plane stress and plane strain in
mid-thickness plates
32 2 Elements of Fractography
where KIC is the plain strain fracture toughness and r0 is the yield stress, therefore
chevrons are not formed in too brittle materials, nor in too thick or too thin plates.
When there are symmetric conditions with respect to a common point of crack
initiation, several sets of chevrons can be formed, all of which will always aim at
the point of fracture initiation.
Beach marks are long, fine and well-defined arch-shaped lines, which go from side
to side on a fractured surface, like the ones shown in Fig. 2.11, and correspond to
the crack propagation front at any given moment. They are formed by changes in
load, environment, or growth mechanism, mainly when a crack stops and re-starts
its propagation. These are also known as “clamshell marks,” or “macroscopic
striations.” It is important not to mistake them for fatigue striations, which are
microscopic.
An additional characteristic of the beach mark is that the arch’s origin point
corresponds to the crack initiation site and the apparent propagation direction of the
set of arches corresponds to the direction of crack propagation. When the beach
mark is not an arch or when the arch´s curving is inverted, it is an indication that the
crack growth rate was not uniform across the front, either due to variation of the
stress levels or fracture toughness in the transverse section of the fractured piece.
Fig. 2.11 Beach marks on a fractured surface. The arrow shows the direction of crack
propagation
2.4 Fracture Lines 33
Fig. 2.12 Modifications of the crack at the arresting and re-starting of crack propagation in a
ductile material, leading to the formation of a beach mark
The formation of beach marks at the arrest and re-start of crack propagation is
explained by the stages shown in Fig. 2.12. In the first row, the crack is closed,
because the load is zero or minimum. When the load increases, as shown in the
second row, the crack tip blunts, because of the plastic deformation, up until a point
when there is a stable extension of the crack, known as a “pop-in,” (third row).
When the instability condition is reached and the crack propagates in a dynamic
way, a step is formed, just as shown in the last row of the figure, which is seen as
the beach mark on the fractured surface.
When beach marks are formed by the arresting and re-starting of crack propa-
gation, in theory, the step height is approximately one half of the CTOD, which is
given by the following equation:
2
4 KIC
CTOD ¼ :
p Er0
So, by measuring the step height of the last beach mark formed just before the
unstable crack propagation, an estimation of the fracture toughness value (KIC) can
be made with the previous equation, provided the material’s yield strength and the
Young Modulus are known, which is relatively easy to do.
34 2 Elements of Fractography
Ratchet marks and ridges are changes in the fractured surface’s level, whose ori-
entation is parallel to the direction of crack growth. The origin of ratchet marks are
the steps formed where multiple cracks, initially on slightly different planes,
interconnect with each other through small secondary cracks perpendicular to the
fracture plane, as shown in Fig. 2.13. Ridges are like small mountain ranges formed
by micro-cracking out of the main fracture plane. Both lines are commonly
observed in cleavage, fatigue or SCC fractures. Normally, the ratchet marks join to
form one single fracture face, making a macroscopic river pattern, while grooves
usually extend without joining (Fig. 2.14).
Multiple cracks at
different levels
Fig. 2.14 Ratchet marks and ridges in a fracture. The block arrow shows the direction of
propagation
2.4 Fracture Lines 35
Shear lips are sharp protuberances formed on the outer edges of ductile fractured
surfaces, caused by high shear stresses in combination with abrupt alterations of the
state of stress towards the end of the fracture. They are of great interest in failure
analysis, because, by being formed in the final stage of fracture (Stage III), they
allow for the identification of the final fracture zone, and therefore the fracture
sequence and location of the initiation point can be determined. In addition, their
height is directly proportional to the fracture toughness, as well as to the material
ductility. The shear lip usually shows a neck, which indicates that the fracture was
due to tensile stresses. Actually, the neck in a tensile test specimen is a shear lip
(Fig. 2.15).
The orientation of the shear lip depends on whether the plastic zone is in plane
stress or plane strain, as is described next. In the condition of plane stress, the
direction perpendicular to the fracture plane corresponds to the maximum principal
stress (direction Y), while the second principal stress corresponds to the propaga-
tion direction (direction X) and the minimum principal stress is in the transversal
direction (direction Z), being of zero value. The Mohr’s circle corresponding to
plane stress is shown in Fig. 2.16. Under this condition, the principal deformations
are all positive (stretching). According to Mohr’s circle of stress, the maximum
shear stress is located on two planes at 45° from the Y and Z directions, thus the slip
plastic deformation will take place on these planes and the shear lip will be parallel
to the propagation direction, as shown in Fig. 2.16. If the plate is too thin, gen-
erally, only one shear lip is formed at 45° in respect to the direction of the maxi-
mum tensile stress.
The plain strain condition, on the other hand, has three positive components of
principal stress, the highest being in the X direction, the medium in the Z direction
and the lowest one in the X direction, and therefore the maximum shear stress will
τ τmax
y τmax
τmax
σ3 = 0 45°
σ1(y) x
σ2 z
Shear 45°
Single
lips shear lip
Fracture plane
Propagation direction
Thin plate
Fig. 2.16 Mohr’s circle, maximum shear stress planes and orientation of shear lips of a ductile
fracture in plane stress
be located at 45° of the X and Y directions, as shown in Fig. 2.17. The shear lip will
then be oriented parallel to the propagation front.
In both cases, planes stress and plane strain, the height of the shear lip is close to
half the size of the plastic zone, which can be estimated by Irwin’s linear elastic
fracture mechanics equation:
1 KIC 2
rp ¼ ;
p r0
where KIC is the fracture toughness and r0 is the yield strength. According to this
equation, the higher the KIC =r0 ration, the higher the shear lip.
Finally, it is important not to mistake the shear lip for the sharp edges that are
frequently observed at the final part of very brittle fractures. Such edges originate in
the change of direction of the fracture propagation path in the last stage of the
fracture process. The easiest way to prevent such confusion is to observe whether
there is a noticeable plastic deformation associated with the edges, that is, if there is
2.4 Fracture Lines 37
τ
τmax τmax
y
45°
σ1 (Y)
x
45°
Fracture surface
Shear lip
Propagation direction
Fig. 2.17 Mohr’s circle, maximum shear stress planes and orientation of shear lips of a ductile
fracture in plane strain (thick plates)
a neck. If there is not a neck and the opposing fracture faces match perfectly, it will
undoubtedly be a brittle fracture.
Fig. 2.18 Macroscopic cleavage in a brittle fracture of a metallic material. The arrows indicate the
crack propagation direction
the direction in which the ranges align, having a greater roughness as the crack
advances to the final separation point.
This type of fracture is formed by exceeding the materials of cohesive strength,
so the fracture orientation is always perpendicular to the maximum tension stress
plane, however, the principal stress direction on the crack´s front may vary as the
crack moves forward, so the cleavage plane may get curved or wavy in that
instance, as in the example given in Fig. 2.19.
Macroscopic cleavage in amorphous materials may result in rough fractures,
with many steps, ridges and ranges and even with fragment detachments in the form
of fish scales, as shown in Fig. 2.20. The direction of propagation corresponds to
the direction in which small ranges align themselves, having greater roughness as
the crack advances forward.
Fig. 2.19 Wavy cleavage plane due to the variation in the orientation of the maximum tension
stress at the front of the crack
2.6 Multiple Cracking 39
Multiple cracking occurs frequently in brittle material fractures when there is a large
enough energy supply. Fracture mechanics energy criterion predicts that under
constant load conditions, the energy release rate (G) increases as the crack grows.
This great availability of energy can favor multiple cracking in Stage III of the
fracture, as shown in Fig. 2.21. The crack starts propagation when G = 2R and its
size is critical.
The formation of crack branching will demand twice as much energy (R) for its
propagation, which is attained when the crack has reached a size of ac þ Da2; G ¼
2R, When the crack’s length extends up to ac þ Da2; G ¼ 3R, there will be enough
energy to get three cracks propagating. This process will not go on indefinitely,
because, as was said before, this occurs only at the fracture’s final stages and before
many branches are formed, thus the piece will be fully separated or broken.
Under impact conditions, on the other hand, the initial energy input is great and
it causes multiple propagations in Stage I of the fracture. However, due to the
system´s demand for energy, multiple cracks will eventually stop and just one of
them will continue propagation. During crack branching, it is common for a crack
to take a Y or T configuration, as shown in Fig. 2.22.
The Y rule says that the crack’s propagation direction is always the direction of
the Y’s branches, as shown in Fig. 2.23.
When a crack intersects with another one at an angle close to 90°, they form a T
type intersection. The rule is that the hat over the T comes first, and then the pole of
the T intersects with the former one, as illustrated in Fig. 2.24. It is also a rule that
the intersecting crack (pole) stops at the intersection.
40 2 Elements of Fractography
G, R
3R
2R
Δa
a Δa1 Δa2
ac
Fig. 2.21 R curve (energy criterion) for crack branching in a brittle material
Fig. 2.22 Multiple cracking on a plate of a pressure vessel that failed by explosion. The arrows
show the propagation trajectory of the cracks
2.7 Microscopic Fractography 41
1. Sectioning. Cut out samples of the fractured piece to a size that fits into the SEM
specimen chamber.
2. Cleaning. Eliminate cutting debris, oil and dirt from the microscopic exami-
nation specimen according to the procedures and recommendations given in
Chap. 1.
3. Specimen preparation. If required, prepare the specimen for observation in the
SEM, by evaporation coating with a noble metal or graphite.
4. Microscopic examination. Once the specimen is in the SEM, first have a look at
low magnification in order to recognize the zones of interest. At this point, it is
advisable to have at hand a macro-photograph of the piece or a drawing,
because, once in the microscope, the change of perspective and contrast (in
SEM, the image is always black and white) may confuse the observer. If the
sample is large enough, or there is more than one sample, a few fine marks can
be scribed in the zones previously identified through the macroscopic exami-
nation, to use them as guidance for quick spotting and as a measuring reference.
The marks can be made using a fine point marker or a needle, and care must be
taken not to make marks on zones of interest. Marks should never be made on
fractured surfaces that are being examined for the first time, those that are
unique or those that are part of a judicial investigation.
5. Photography. Once the area to be observed has been selected, begin at low
magnification and increase amplification up to a level that allows for clear
identification of the desired features. Here, it is recommended to pick out a
single area and increase amplification in sequence, from 100, 500 and
1000. Greater amplifications will be decided according to the level of reso-
lution needed to reveal fine fractographic features. It is also recommended that at
least one shot be taken with low, middle and high amplifications of the zone of
interest.
6. Representativeness. An important recommendation is that attention should be
focused on those characteristics that appear more often on the fractured surface,
with little to no attention being paid to rather small or rare details that might be
attractive, but are not representative of the main fracture mechanism. The
researcher must bear in mind that he or she is just observing a little fraction of a
whole, and there is always the risk of making incorrect generalizations.
Likewise, it is not advisable to take photographs of every detail on the fractured
surface, because, in addition to making the examination longer and more costly,
the investigator may end up tired and confused when preparing a written report.
7. Micro-Analysis. Once the fractographic observations and their photographing
have concluded, the investigator can proceed to carry out both the chemical and
physical analyses required. It is advisable to do this at the end of the exami-
nation, because occasionally, the analysis techniques may damage the surface or
alter the material itself.
8. Preservation. The final step in the microscopic examination of fractured surfaces
is the preservation of the specimens, including parts and fragments that were not
examined.
2.8 Fracture Micromechanisms 43
Grain
boundary
Grain
Intergranular
Crack
Transgranular
Non-crystalline Crack
Crack
Transgranular
Crystalline
Fig. 2.26 Process zone of a crack in a ductile material, seen at low magnification in an SEM. The
process zone (PZ), the stretching zone (SZ) and the slip lines (SL) are clearly visible
Atomic bond
Stress
rupture
normal to
the plane
Cleavage
plane
New
surfaces
Decohesion
Grain boundary
(weak interface)
plane will be the plane perpendicular to the maximum principal stress, whereas in
crystalline materials, the cleavage plane is a specific plane family for each type of
crystalline structure. The specific cleavage plane is the one that is best oriented in
respect to maximum principal stress. Figure 2.27 shows a schematic model of this
mechanism. As expected, this mechanism does not involve plastic deformation,
therefore it is predominantly brittle. In addition, in polycrystalline materials, the
fracture is transgranular and its macroscopic appearance is faceted and grainy.
Decohesion- This is the direct separation of any type of interface in the material.
It differs from cleavage because the fracture does not occur in a crystalline plane,
but the separation rather takes place on the interfaces, whose cohesive stress has
been surpassed by a normal stress component or it has been weakened by defect
build-up or by accumulation of chemical species. This mechanism often occurs on
the grain boundaries because of their high diffusivity and solubility, so the deco-
hesion fractures in polycrystalline materials are intergranular. This mechanism is
usually brittle, since it requires little or no plastic deformation (Fig. 2.28).
46 2 Elements of Fractography
Plastic Flow- At the tip of the crack or at the root of a sharp notch, under tension
stress (or Mode I fracture), a crack can propagate due to the plastic deformation at
the crack tip or notch root. The state of stress in these flaws generally causes the
maximum shear stress planes to be in an inclined direction with respect to the
fracture plane, so the plastic flow will be in those directions, displacing material
away from the crack tip, and thus causing the crack to move forward, just as shown
in Fig. 2.29. Plastic deformation can occur by dislocation slip, diffusive flow or
viscous flow. In this mechanism, the fracture will be transgranular and
non-crystalline, and its macroscopic appearance will be fibrous and opaque.
Ductile shear- This fracture mechanism occurs when the shear strength of the
material is surpassed by the maximum shear stress, so a sliding separation of the
maximum shear plane occurs. Obviously, this mechanism appears only in shear or
torsional fractures of ductile materials (Fig. 2.30).
Direction of maximum
Plastic flow
shear stress
zone
Plastic flow
zone
Crack extension
Crack Void
nucleation
Crack path
Process zone
Growth and
coalescence of voids
Crack extension
Nucleation, growth and coalescence of voids- This type of fracture occurs when
micro-voids are formed in the core of the material or some interface. The voids
grow under the action of stress until they join and coalesce, thus creating a
macroscopic crack. The coalesced voids may also connect to another crack, causing
it to grow (Fig. 2.31).
Chapter 3
Brittle and Ductile Fractures
Abstract In this chapter, the main characteristics of brittle and ductile fractures,
caused by single load application, are described. Each section begins with a
description of the macroscopic features that allow for identification of brittle and
ductile fractures, and continues with the description of the mechanisms of formation
of the fractographic features of each type of fracture. Examples of each described
feature are given, as well as the particular aspects of brittle and ductile fractures in
non-metallic components.
3.1.1 Cleavage
Cl ea va g e
p la ne
Grai n
Fac et b y Cl ea va g e
clea va ge p la ne
of t he
gra i n Grai n
brittle way, also show cleavage in a practical way. However, in this case, the
fracture plane is defined simply by the direction of the maximum principal stress.
The cleavage resistance is determined by the cohesive strength between the
atomic planes of a solid body. In the absence of stresses by external loads, the
atoms are at an equilibrium distance, where the inter-atomic attraction and rejection
forces are at equilibrium. If the intention is to get the atoms at a distance shorter
than that of equilibrium, a rejection will be experienced, while, on the other hand, if
they are separated, there will be an opposing attraction force. Figure 3.2
schematically shows the variation of the inter-atomic force with respect to the plane
separation distance.
The attraction force shows a maximum, which, if surpassed, will cause the
atoms’ separation, thereby making it the theoretical cohesive strength of the
ao
Rep u l sio n
3.1 Brittle Fracture 51
material. When plane separation happens in a perfectly brittle fracture, the work
applied will be transformed into surface energy. The next expression represents the
theoretical strength in terms of surface energy ðcs Þ; the elastic modulus ðE Þ and the
equilibrium inter-atomic distance, that is, the lattice parameter ða0 Þ:
rffiffiffiffiffiffiffiffi
EcS
rTEO ¼
ao
which is about 8 times higher than the experimentally measured maximum strength
of any material (approx. 3 106 psi). This simple calculation indicates that there
must be something that reduces the material strength so that the cohesive strength
approximates to the experimental values.
A possible cause of the reduction of the cohesive strength is the existence of
small flaws that concentrate stress inside the material, such as micro-cracks. The
calculation of the flaw size that concentrates stresses at levels of cohesive strength
can be done as follows: Consider an infinite plate under a uniform tension stress
ðrÞ; which contains an elliptical hole of length equal to 2c, like the one shown in
Fig. 3.3.
The maximum stress ðrMAX Þ at the ends of the hole is given by
rffiffiffi
c
rMAX ¼ 2r
q
MAX MAX
2c
52 3 Brittle and Ductile Fractures
If the flaw becomes sharp enough, the ratio at the tip ðqÞ gets closer to zero, so
the maximum stress increases until it equals the theoretical cohesive strength, thus,
equating the expressions for theoretical strength ðrTEO Þ and maximum stress
ðrMAX Þ and
solving for the applied stress, the fracture strength by cleavage is
obtained rCleavage ; and is given by
rffiffiffiffiffiffiffi
Ecs
rCleavage ¼ :
4c
In 99.8% density Alumina (typical material that fractures in the brittle fashion),
the following data are known: rCleavage ¼ 30 103 psi, E ¼ 56 106 psi, cS ¼
103 in. lb=in.2 . Solving for the crack size: 2c ¼ 3:1 105 in:; about 0.8 l. A flaw
of such dimensions is likely to exist in a piece of Alumina, so it can be concluded that
cleavage fracture can be promoted by micro-cracks that may occur naturally in a solid
material.
The direct separation of cleavage planes produces a transgranular fracture with
facets that are smooth and shiny. Because the facets are oriented at different angles,
the fracture looks grainy and produces specular reflection when observed under
direct light at the macroscopic level, as shown in Fig. 3.4.
At the macroscopic level, cleavage fractures have a facet-like appearance, which
is very easy to recognize, as shown in Fig. 3.5. The passing of the fracture through
the grains creates a set of fine steps on the facet, known as the “river pattern.” It is
so called because it looks like a river whose branches join with the main stream as it
goes downhill, so to speak. The river pattern branches join in the growth direction,
reducing in number as the crack front moves away from the grain boundary where
they were initially formed.
The origin of the river pattern is due to the interconnection of cleavage planes
with different orientation as the crack moves across the grains, where the cleavage
plane has to rotate to keep the continuity of the propagating plane. Since the
rotation cannot be done abruptly, it is done by little steps whose edges make the
river pattern, as schematically illustrated in Fig. 3.6.
The shape of the river pattern allows us to determine the microscopic direction of
a crack within a grain, as explained earlier (it is downstream as the branches join
together). However, when observing cleavage in a group of grains, like those in
Fig. 3.5, it will be noticed that some of the grains have a growth direction com-
pletely opposite to the microscopic growth direction. This is because some grains
ahead of the crack tip may fracture before those that are connected to the main
crack. These grains may be weak or favorably oriented, so the tension field ahead of
the crack is high enough to cause their cleavage, as illustrated in Fig. 3.7. However,
despite these events, the average growth direction, resulting from the vectorial sum
Ro t at io n a n g le
bet we e n gr ai ns ( ) Ste p s to a cco m mod at e t he
ro tat io n b e t we e n gr ai n s
Grain 2 Ri v er
patt er n
fo r me d b y
th e fra ct ur e
Grain 1 Cra c k
st ep s
d irec tio n
Cl ea va g e
p la ne s
High stress
process zone
Fig. 3.7 Cleavage of a grain in a direction opposite to the main crack growth direction, due to the
stresses in the process zone that are high enough to fracture weak or favorably oriented grains
of individual growth components of every grain, will match the macroscopic crack
growth direction.
When cleavage takes place in brittle precipitates inside the grains, the fracture of
the precipitate occurs almost free of the river pattern, but once the crack goes
through, a river pattern will develop, as shown in the example in Fig. 3.8.
Another characteristic of cleavage fractures is a set of fine steps, formed when
the main fracture plane crosses zones with high dislocation density, as occurs in
heavily cold formed metals. These fine steps create an array known as “feather
marks,” like the one shown in Fig. 3.9.
3.1 Brittle Fracture 55
Fig. 3.9 Fine steps forming feather marks in a cleavage fracture of a heavily cold formed metal
Cl ea va g e
p la ne
3.1.2 Quasi-Cleavage
Cl ea va g e P las ti c
p la ne defo r mat io n
Cra c k
Ste p
Fig. 3.12 Quasi-cleavage, caused by plastic deformation at the steps of a brittle fracture
A quasi-cleavage fracture is rougher and has coarser river marks than those of pure
cleavage, but it still keeps the facets, as seen in Fig. 3.13. Quasi-cleavage is a very
common type of fracture in impact-fractured steels at temperatures close to the
ductile-brittle transition, although its identification is difficult, as it may be mistaken
for fatigue or environmental fractures.
3.1 Brittle Fracture 57
Fig. 3.13 Quasi-cleavage in steel fractured by impact, close to the ductile-brittle transition
temperature
3.1.3 Pseudo-Cleavage
In near threshold fatigue fractures and some forms of stress corrosion cracking,
faceted fractures having well defined river patterns may develop. Such forms of
fracture, despite showing characteristics identical to cleavage, are not cleavage
fractures. In the case of stress corrosion cracking, facets are developed by prefer-
ential plane dissolution or sensibilization and decohesion of inter-faces, and in the
case of fatigue, the facets correspond to the slip planes in which the crack propa-
gates by dislocation cross-slip.
The pseudo-cleavage fracture mechanism will be discussed in the next chapter.
At this point, it is enough to say that pseudo-cleavage and cleavage fractures are
quite similar, as shown in Fig. 3.14. In practice, the best way to distinguish between
pseudo-cleavage and cleavage is through macroscopic examination, where it can be
determined if a fracture is the result of fatigue, and so the presence of
pseudo-cleavage can be anticipated. Another indication of pseudo-cleavage is the
presence of well-defined slip lines around the crack on a polished surface, as will be
seen in Chap. 4.
pattern. The origin of the river pattern in amorphous material fractures is the same
as in polycrystalline material cleavage, that is, it is due to rotation of the fracture
plane, and the only difference is that the fracture rotation is due to variations of the
stress field orientation.
Along with the river pattern, there are two other characteristics that can be found
in cleavage fractures of amorphous materials. One is the “teardrop mark,” which is
formed when the crack front meets a particle or inclusion that temporarily alters
both stress fields. The second is the Wallner lines; these are waves on the fractured
surface that look similar to beach marks. They are formed by variations in the
orientation of the stress field in front of the crack tip during crack propagation. An
example of a macroscopic cleavage fracture in an amorphous material is shown in
Fig. 3.15.
Fig. 3.16 Intergranular brittle fracture in steel. Left, crack path as seen in a metallographic
specimen. Right, as seen by SEM
A ductile fracture is one with previous plastic deformation, visible in most parts of
the body volume. This type of fracture has been widely studied in tension tests and
its typical sequence is:
i. Uniform deformation.
ii. Localized deformation, with the formation of a diffuse neck.
iii. Beginning of instability, with the formation of a localized neck.
iv. Final fracture by tearing shear.
Stage ii corresponds to the maximum stress on the stress-strain curve, and it
occurs when the material has exhausted its capacity to accommodate uniform
deformation; then, a constriction called a neck takes place in the transversal section.
The neck curvature induces a triaxial stress state, which, in turn, leads to strain
localization within a narrow zone, forming a sharp radius neck. In this zone, the
final fracture by tearing shear takes place. For this reason, the final portion of a
ductile fracture is a shear lip with slant sides. The latter sequence produces the
typical uniaxial tension fracture known as cup and cone. This mechanism is shown
schematically in Fig. 3.17.
As described above, a ductile fracture in uniaxial tension produces the cup and
cone shape, as shown in Fig. 3.18.
It is interesting to note that the macroscopic geometry of ductile fractures, and
particularly the form of the shear lip, depends on the state of stress, as will be seen
below.
The ductile fracture of plates in tension may take place in one of two extreme
cases:
Cone
Cup
Fig. 3.18 Macroscopic appearance of a ductile fracture in uniaxial tension. Left, cup fracture.
Right, cone fracture
(a) Plane stress- In this case, the components of the maximum shear strain are on
planes transversal to the fractured surface. This produces two necks at each
crack side and the fracture will show a shear lip parallel to the crack growth
direction, as shown in Fig. 3.19a.
(b) Plane strain- Here, the maximum shear strain components are on inclined
planes in front of the crack tip. This produces a neck parallel to the front of the
crack and the fracture will have a neck parallel perpendicular to the crack
growth direction, as shown in Fig. 3.19b.
In the case of very thin plates, the fracture may be fully controlled by the shear
strain and the orientation of the fracture plane will correspond to the orientation of
the maximum shear plane, thus creating a fracture with one or two sides inclined at
about 45° from the tension direction, as shown in Fig. 3.20.
Plastic
zone Plastic
zone
Neck Neck
Crack
Crack
Fig. 3.19 Neck shape and location on a ductile fracture of plates fractured in tension
62 3 Brittle and Ductile Fractures
+ 1
3 =0 + 2
When the ductile fracture is pure shear, the fractured surface will be on a plane
parallel to the maximum shear stress direction and the shear lips will not form. The
most common pure shear fractures are torsional fractures of cylindrical bars, in
which case the fractured surface is as shown in Fig. 3.21.
3.2 Ductile Fractures 63
The basic mechanism of ductile fracture is void nucleation and growth, which
consists of the following: once a neck is formed, a triaxial stress state is created at
the mid-thickness section of the solid body. This triaxial stress induces the nucle-
ation of voids, usually at inclusions or second phase particles. The voids grow by
plastic deformation until the ligament between the neighboring voids becomes
small enough to produce their coalescence, forming an internal cavity. In that
moment, the shear stress in the remaining transversal section (ligament) increases
until a failure by sliding shear occurs. This mechanism is shown in Fig. 3.22.
The voids causing the ductile fracture mechanism are of microscopic size and
look like little craters or dimples on the fractured surface. The shape and size of the
dimples are directly related to the size, shape and distribution of the inclusions or
second phase particles that are their nuclei, as well as to the mode of loading, as will
be described in the next paragraphs.
Effect of particle size: The rule is that fine and closely spaced particles produce
small dimples, because the voids grow a little before coalescence; however,
whereas coarse particles produce large dimples, the final size of the dimple will
depend on the particle spacing and the material’s ductility. Normally, metals have a
combination of large and small particles, therefore it is common to observe a
mixture of dimple sizes, as seen in Fig. 3.23.
Effect of particle shape: Spherical particles produce more or less equiaxial
dimples, whereas long particles produce long and narrow dimples, like the ones
shown in Fig. 3.24.
Triaxial
stress
max
Void
Neck
Fig. 3.22 Schematic representation of the ductile fracture mechanism. i Void nucleation, ii void
growth and iii final separation by plastic shear
64 3 Brittle and Ductile Fractures
Fig. 3.23 Microscopic view of dimples in a ductile fracture under uniform tensile loads. Notice
that larger dimples are produced by larger particles, and vice versa
Fig. 3.24 Long and narrow dimples produced by long nucleating particles
Effect of ductility: High ductility materials produce deep and narrow dimples,
whereas less ductile materials produce shallow and wide dimples. Figures 3.25 and
3.26 shows an example of each.
3.2 Ductile Fractures 65
Effect of load application mode: Tensile loads uniformly distributed across the
transversal section will produce equiaxial dimples, whereas lateral or tangential
loads will form parabolic dimples, as shown in Fig. 3.27.
66 3 Brittle and Ductile Fractures
Fig. 3.27 Ductile fracture dimple shape according to the loading mode
Void Void
nucleus nucleus
Distorted
Particle crystal
Particle decohesion
cleavage
Void nucleus
Fig. 3.28 Void nucleation mechanism. a Fracture of particles, b decohesion of interface and
c dislocation pile-up
68 3 Brittle and Ductile Fractures
Vo id w a ll
Slip steps
S l ip b an d s
in tension is more than 5%, the ductile fracture characteristics, such as necks and
shear lips, can be easily observed, so this may be a good criterion for separating
brittle from ductile fractures.
In ferrous materials, brittle fracture occurs at service temperatures below a
specific value known as ductile-brittle transition temperature, which is obtained
from the Impact Energy ðCvÞ versus Temperature ðT Þ curves, like the one shown in
Fig. 3.30.
The ductile-brittle transition is located in the mid-part of the Cv versus T curve,
just when brittle fracture percentage is 50% (the measurement is done on the
Charpy bar fracture surface by fractographic techniques). This criterion is mostly
correct, because the quantification of brittle fracture is not always precise. For this
reason, threshold impact energy values are more frequently used to establish the
onset of brittle fracture or unacceptable toughness values. For example, in the
pipeline industry, steels with Cv levels below 20 ft lb at 0 °C are considered too
brittle and their use is restricted to temperatures in the upper shelf region of the
curve.
D uct il e -B ri tt le
tra n si tio n
Te mp erat ur e [ ° C]
3.2 Ductile Fractures 69
Fig. 3.31 Combination of fracture mechanisms at the microscopic scale: ductile fracture by
micro-voids, combined with brittle fracture cleavage and intergranular fracture
Abstract In this chapter, the general aspects of fatigue fracture, such as fatigue life,
the three stages of fatigue and the mechanical characteristics of materials’ behavior
under cyclic loads, are presented. Once the fatigue phenomenon is understood from
the mechanical point of view, the fractographic characteristics of fractured surfaces
at each one of the three stages are described in detail. The mechanisms of fatigue
crack nucleation and propagation are described, along with the typical macroscopic
and microscopic features of fatigue fractures. Emphasis is placed on the description
of the origin and characteristics of microscopic fatigue striations, as one of the most
distinguishable aspects of fatigue fractures. A description of the fractographic
characteristics of Stage II fatigue fractures in the absence of striations is also
included. The chapter finishes with a description of corrosion-fatigue fractures.
Shear lip
Final fracture
(Slant fracture)
(rough)
(A) Elastic strain: The entire body is elastically strained, and there is only a small
plastic zone at the crack tip. As the stress level is relatively low, a high number
of cycles is necessary to reach fracture (it might be several million), and
therefore it is known as high cycle fatigue.
(B) Elastoplastic strain: There is plastic strain in a large portion of the ligament
(remaining transversal section in the crack plane), so the crack propagates
across a plastically deformed region. In this case, fracture takes place within
thousands or hundreds of cycles, so it is called low cycle fatigue.
(C) Fully plastic strain: The entire volume is under cyclic plastic strain, so the
fracture occurs within a few cycles, and is called very low cycle fatigue.
74 4 Fatigue Fracture
The first case (elastic strain) is the most interesting in research and practice,
because all structural and mechanical components prone to fatigue are designed to
operate under an elastic strain regime. When cracks caused by fatigue appear in
these components, the conditions of linear elastic fracture mechanics (LEFM)
are satisfied, so the fatigue crack growth can be characterized by the stress intensity
factor amplitude (DK).
The shape of the load cycle is the first major mechanical variable of fatigue, as it
determines the magnitude and strain rate in the crack tip. As mentioned earlier, the
load can be either fluctuant, cyclic or random, as shown in Fig. 4.2.
In the three stages of fatigue, the crack growth rate depends on the stress
amplitude at the crack tip, DK, which, under elastic linear conditions, is determined
as follows:
The general equation for K, according to LEFM, is
pffiffiffiffiffiffi
K ¼ Pb pa;
DP ¼ Pmax Pmin :
Thus,
pffiffiffiffiffiffi
DK ¼ DPb pa:
L
o
a
d Time
½ frecuency
Kmax
K
Δ K = K máx - K mín
Kmin
The load cycle is completely defined by the load ratio R, which is defined as
R ¼ Pmin =Pmax :
Based on the R value, the following load cycle types are defined. Note that the
sign of R is defined by the sign of the applied loads and not by the K values, since
there are no negative values of K (Table 4.2).
During Stage I, the crack usually initiates in a stress concentrator, which, in a
generic way, is a notch of radius q. It has been found that the ratio DK/q controls
the number of crack initiation cycles Ni. The limit below which fatigue cracks do
not grow (DKth), and therefore the fatigue life of the component is infinite, is related
to both the yield strength r0 and the notch root radius of the stress concentrator by
the following empiric relation:
pffiffiffiffiffiffiffiffi
DKth ¼ 10 qro :
In Stage II, at constant load amplitude, as the crack propagates, the stress
intensity factor increases, as does the crack growth rate. This process goes on until
the maximum stress intensity factor value equals the material fracture toughness KIC
and the final fracture occurs, meaning the crack growth rate (da/dN) is a function of
DK. The Fig. 4.4 schematically shows this behavior.
Paul C. Paris, in 1962, proved that in a logarithmic plot of da/dN versus DK data,
there are three well-defined regions that correspond to the three stages of fatigue
ΔK a
Δ P = Cte.
da / dN
a N
crack growth, as shown in Fig. 4.5; he also found that in Stage II, the following
relation is valid:
da
¼ C DK m ;
dN
where C and m are empiric constants. This equation has been fundamental for
fatigue research and has allowed us to analyze the effect of multiple factors, both
internal and external, in fatigue crack growth. These studies have shown that the
most important influence comes from the microstructure and the environment.
Figure 4.5 summarizes such influences.
Stage II
Stage I KMAX = KIC
Paris region
Near threshold
.Crystalline
. Transgranular
fracture. fracture
.Effect of . Influence of
microstructure environment
Stage III
Unstable
da/dN . Combination
with static
modes of
da/dN = C ΔKm fracture
Fatigue threshold
Δ Kth
ΔK
Fig. 4.5 Fatigue crack growth rate behavior based on the intensity stress factor amplitude (Paris
plot). Logarithmic scale
4.1 General Aspects of Fatigue 77
Stress
Monotonic
strain
Cyclic Cyclic strain
plastic zone loop
Strain
Crack
Monotonic Δε plastic
plastic zone
Δε elastic
Fig. 4.6 Cyclic plastic and monotonic zones at the tip of a fatigue crack and the strain-stress cycle
in both zones
From the mechanical behavior point of view, the most important zone of a
fatiguing component is the plastic zone, initially generated by the stress intensifi-
cation at the crack tip. Due to the fluctuant nature of stresses, the plastic zone is
divided into two regions, as shown schematically in Fig. 4.6. The smaller region,
located right in front of the crack tip, is a cyclic strain zone, where strain goes from
tension to compression, due to the change of direction of the load during each cycle.
The cyclic strain zone is surrounded by a larger strain zone, where the deformation
is monotonic, and its size depends on the maximum stress value.
The crack propagation basically depends on the behavior of the cyclic zone, but
it is strongly influenced by the size of the monotonic zone. Paris demonstrated that
the cyclic plastic zone size (rc) is approximately
1 DK 2
rC ¼ ;
8p r0
where ΔK is the stress intensity factor amplitude and r0 is the yield strength. The
size of the monotonic plastic zone can be calculated according to the Irwin cor-
rection formula
1 Kmax 2
r0 ¼ :
p r0
Based on the aforementioned formula, the cyclic zone is about eight times
smaller than the monotonic zone. The actual size of the cyclic zone, of course, is
affected by several factors, like strain hardening, anisotropy and strain rate, among
others.
78 4 Fatigue Fracture
At the macroscopic level, the crystalline nature of Stage I mechanisms leads to the
formation of faceted fractures with dense and well-defined river patterns, as shown
in Fig. 4.8. This type of fracture is called pseudo-cleavage, due to its similarity to
cleavage. Pseudo-cleavage can be very hard to differentiate from cleavage, so a
novice fractographer may formulate an incorrect interpretation. In order to prevent
confusion, it is recommended that a detailed macroscopic inspection of the frac-
tured component be carried out before proceeding to the microscopic observation.
If the latter is not possible, the safest way to identify pseudo-cleavage is by
carrying out a crystallographic trace analysis to identify the planes and directions of
the river pattern edges, which should correspond to the slip system of the particular
fatigued alloy, and not to cleavage systems. The same analysis can be done on the
slip lines that appear on the free surface of the piece, as long as it has a mirror-like
polish, as shown in Fig. 4.9.
In Stage I fatigue of polycrystalline material, a common characteristic of the
fractured surface is the formation of a high number of ridges parallel to the crack
propagation direction. These marks are formed due to the presence of a component
of Mode II displacement in the crack displacement, especially in planar slip
materials. Figure 4.10 shows an example of this. These fractures are difficult to
interpret, and the best recommendation, again, is to carry out a good macroscopic
examination, to make sure that it is a Stage I fatigue fracture.
In pieces with neither pre-existing cracks nor stress concentrators, the basic
nucleation and slow propagation fatigue crack mechanism is by dislocation slip;
Fig. 4.8 Pseudo cleavage on a single crystal of a nickel base super alloy, fatigued in air at room
temperature
80 4 Fatigue Fracture
Fig. 4.9 Slip lines around a pseudo-cleavage fracture on a nickel alloy single crystal fatigued in
Stage I
Fig. 4.10 Stage I fatigue fracture in a polycrystalline material. Stainless steel fatigued in air at
room temperature
nonetheless, when there are either pre-existing cracks or sharp stress concentrators,
Stage I is suppressed.
The most widely accepted model of fatigue crack nucleation is known as the
intrusion-extrusion mechanism, proposed by Wood. In this model, the dominating
slip plane goes through an alternated slip process that leads to the formation of
4.2 Fractography of Stage I Fatigue 81
Free
surface
Intrusion Nucleated
crack
Slip plane
Extrusion
Back and
forth slip
Sharp
Free surface extrusion
Decohesion of a slip
Nucleated
band
crack
Fig. 4.12 Forsyth’s model for fatigue crack nucleation in ductile metals
Dislocation
line
Fig. 4.13 Mott’s double cross-slip model of crack nucleation in Mode II fatigue
Absorption and dissolution of gases. The intense activity of slip bands on a free
surface during fatigue, promotes that the low molecular weight gaseous species
penetrate into the metal through the slip bands. Two phenomena may occur: one is
decohesion along the slip bands and the other, more visible one, is the formation of
micro-cavities that, when they coalesce, originate a crack (Figs. 4.15, 4.16 and
4.17).
Formation and rupture of an oxide layer- Under highly oxidizing conditions and
in dry environments (high temperature), thick and brittle oxide layers can be
formed. When fractured by tension, they originate cracks that, combined with
dislocation bands, may provide high rate diffusion paths for oxygen, so the slip
band becomes internally oxidized, and subsequently fractures, causing the crack to
grow. This type of fracture is also known as thermal fatigue.
Slip localization- In this model, the absorption of hydrogen from atmospheric
humidity dissociation promotes the localization of a dislocation slip, due to the
4.2 Fractography of Stage I Fatigue 83
Fig. 4.14 Ejection of a metal fragment from a step in a crystalline fracture of a nickel base
superalloy single crystal fatigued in vacuum
Gas diffusion
Oxide layer
Oxide layer cracking Crack
Slip band
Oxidación
interna
Fig. 4.17 Thermal fatigue crack in steel caused by the rupture of oxide intrusions
Fig. 4.18 Typical characteristics of a Stage II fatigue fracture: smooth surface covered by beach
marks
polymers, as shown in Fig. 4.19, where the typical features of fatigue can be clearly
observed.
The fractured surface extension in Stage II depends on the maximum stress and
the stress concentration level, as explained. The most common type of fatigue is
simple bending of beams, bars and shafts. In this type of load, the piece is bent
86 4 Fatigue Fracture
Low
High
Fig. 4.20 Effect of stress amplitude and stress concentration level on the appearance of bending
fatigue fractures. Kt is the stress concentration factor
during the first half of the cycle and then the load is released. Since the outer radius
of the bent bar is in tension, the cyclic load is tension-tension and the crack will
start at the surface of the outer radius, because that is where the highest tensile stress
is located. As the crack propagates towards the center, the neutral axis moves
forward, until the critical crack size is reached and the piece fails. One special case
is rotation bending: in such a case, there are two initiation sites located at opposite
sides of the piece, so the final fracture zone will be at the middle section of the
piece. Figure 4.20 schematically shows the appearance and extension of the fatigue
fractured surface of a round bar in simple or rotation bending as a function of the
stress concentration level and the load amplitude.
In alternating torsion, the maximum tension stresses are at ±45° from the lon-
gitudinal shaft axis, therefore two groups of cracks will be formed in planes at 45°
from the shaft axis and will be perpendicular to each other, forming a star-type
fracture, as shown in Fig. 4.21.
Torsion
Direction of σmax
+45
Long. axis
-45
Cracks
spacing is close to the macroscopic crack growth rate, thus it is believed that each
striation corresponds to one load cycle. The presence of striations indicates, without
doubt, that the observed fracture was caused by fatigue, but their absence does not
mean the opposite, because striations may fail to form under a variety of conditions.
Depending on their appearance, the striations are classified into two groups:
1. Ductile striations: The striation profile is wavy and smooth, as in Fig. 4.22.
2. Brittle striations: The striation profile is irregular or saw tooth-like.
Occasionally, they look like micro-cracks, as shown in Fig. 4.23.
88 4 Fatigue Fracture
Crack tip
Crest Valley
crack propagation front and cause the crack to locally accelerate or retard, while the
rest of the crack front grows at a uniform rate in a single phase zone.
A generally accepted observation is that striation spacing matches the macro-
scopic crack growth rate (da/dN). This has not been fully confirmed, mainly
because, at slow growth rates (less than 10−6 mm/cycle), the resolution limit of the
scanning electron microscopes does not allow for measuring the striation spacing.
Table 4.3 allows for a better appreciation of the relation between striation spacing
and the magnification required for their observation.
Figure 4.26 shows the relation between striation spacing measured in an MEB
and the macroscopic crack growth rate da/dN. It can be seen that striation spacing is
constant below da/dN values of around 10−5 mm/cycle, but this has to do with the
electron microscope resolution used for studying these fractures. In order to observe
Table 4.3 Fatigue crack growth rate and amplification needed to observe striations (in case they
do exist)
Macroscopic crack Typical Magnification necessary for Observation technique
growth rate scale observation
10−3 mm/cycle 1l 100 Optical microscope (poor
resolution)
10−4–10−5 mm/cycle 0.1 l– 1000 a 10,000 Scanning electron
10 nm microscope (SEM)
10−6–10−7 mm/cycle 5A Approx. 150,000 High resolution SEM or
TEM
<10−7 mm/cycle 1A 1,000,000 Tunnel effect or atomic
force
90 4 Fatigue Fracture
10-5
da/dN macro
log ΔK
So far, a unique mechanism for fatigue striation formation has not yet been
recognized, not even for Stage II crack growth. In general, it is accepted that the
cracking mechanism requires both some crack tip opening and slip plastic strain at
the crack tip flanks, however, it is not clear whether the presence of a corrosive
environment introduces a decohesion or cleavage component, nor is it clear how
fatigue crack growth takes place in inert or vacuum environments. This makes the
proposed mechanism models simple and based on direct observations of the
propagating crack tip.
The most representative model of striation formation is the so-called reversed
slip mechanism, as proposed by Pelloux in 1969. It considers that a crack moves
forward by alternate plastic slip at the flanks of the crack tip, so the striations are
formed in air because of the lack of slip reversibility induced by surface oxidation.
This model explains the existence of striations in air and their absence in vacuum,
and it is consistent with the experimental evidence, as seen in Fig. 4.27. Currently,
it is the most accepted model, although the explanation that an oxide layer formed
at the crack tip is strong enough to affect slip reversibility is still questionable, since
in noble metals that show striations, the oxide layer is very thin.
Nix and Flower observed thin foils of fractured surfaces extracted from alu-
minum fatigued in air, in TEM at 1 MV, finding alternating bands of high and low
dislocation density, matching the striation spacing. They assumed that during crack
extension, there is a cleavage component caused by hydrogen embrittlement of the
plastic zone. The hydrogen is generated by the decomposition of humidity in the
atmosphere.
McEvily and Gonzalez gathered evidence of the main effects of air on fatigue
crack propagation, as compared to propagation in vacuum environments. Such
effects are: faster crack growth rates, higher fracture roughness, less crack tip
opening displacement (CTOD) and plastic strain localization. These even when the
calculated plastic zone size is the same in air and vacuum. They observed that the
crack tip stretching is greater and the deformation more widespread in vacuum than
in air.
Fig. 4.27 Crack tip of 304 stainless steel fatigued in vacuum (left) and in air (right) at the same
DK level
92 4 Fatigue Fracture
Figure 4.26 shows the fatigue crack tips of stainless steel fatigued at the same
DK level in air and vacuum, respectively, where it can be seen that the crack tip is
sharper and the plastic zone is smaller in air as compared to vacuum.
As mentioned earlier, striations are not always formed on Stage II fatigue fractured
surfaces. Some of the conditions under which striations are not formed are:
1. Vacuum or inert environments.
2. High strength materials; for example, quenched tempered steel with hardness
above 30 HRC.
3. Aqueous environments and highly oxidizing or high temperature air. Under
these conditions, it is common to observe striation-like marks, but their spacing
does not match the macroscopic crack growth rate. These have been called
“ill-defined striations.”
4. Creep fatigue interactions.
5. Low DK values. For example, in stainless steel at DK < 30 MPa√m, striations
are not formed (da/dN = 10−5 mm/cycle).
6. Near DKth values.
7. Mode II and III fractures.
The fractography characteristics under these conditions are usually varied and
complex, and therefore difficult to interpret. In steels fatigued in air and low DK, the
fractured surface has a smooth appearance, with small ridges aligned in the
direction of crack growth, forming microscopic river patterns. Figure 4.28 shows
two examples of fatigue fractures without striations.
Fig. 4.28 Fatigue fractured surfaces without striations. Type 304 stainless steel, fatigued in
vacuum (left) and near threshold fatigue fracture un carbon steel (right)
4.4 Corrosion-Fatigue 93
4.4 Corrosion-Fatigue
da/dN da/dN
Dissolution
control
Mechanical fatigue
(Inert environment)
ΔK ΔK
MATERIAL ENVIRONMENT
Environmental
fracture
LOADS
Corrosion is the oxidation reaction of a metal resulting from interaction with the
surrounding environment. The metal oxidation reaction is
M ! M n þ þ ne :
DG\0;
where DG is the Gibbs free energy. If DG > 0, the reaction occurs in the opposite
way and there is no corrosion. Since the reaction is an oxidation, there must be a
reduction reaction that absorbs the electrons released by the oxidation reaction. The
partial reactions of oxidation-reduction are
where M is a metal (the oxidized element) and X is the oxidizing agent, and
therefore it is the element to be reduced and ne− is the number of the transferred
charge.
The anodic and cathodic reactions take place on physically separated areas of the
free surface exposed to an electrolyte (which is part of the environment). These
areas are electrically connected by the metal, as seen in Fig. 5.2. The ion transport
is carried out in the electrolyte. The reduction reaction requires the transport of the
oxidizing agent from the electrolyte to the metal surface, so the diffusion of
chemical species is necessary. Such diffusion is through the layer closest to the
surface and is usually a factor that controls the reaction rate.
As shown in Fig. 5.2, in the anodic region, the metal is oxidized and becomes
part of the environment, causing the characteristic corrosion metal loss. Oxidation
means electron loss, so the lost electrons travel through the metal to the cathodic
region, where they are taken by the oxidizing agent and the reduction reaction takes
place. The reduction reaction is electron gain; therefore, the metal keeps its metallic
state and the cathodic zone will not corrode.
As time passes, the cathodic region eventually becomes anodic and will start to
dissolve, while another zone becomes cathodic to balance the reaction. When this
mechanism of cathodic to anodic area alternation occurs continuously, a fairly
uniform layer of the corroded piece is lost, and thus this is called “uniform cor-
rosion,” even though, at the microscopic scale, the surface shows numerous dimples
5.2 Fundamentals of Metal Corrosion 101
Oxidizing agent
ne- are transport Cathodic
conducted to the region
(+) X X n-
Metal Electrolyte
Anodic
ne-
region
given
Oxidized metal to
M M n+
the electrolyte
Cathodic Anodic
Cathodic Anodic Cathodic
zone zone
zone zone zone
Uniform
metal loss Localized corrosion
Surface
roughness
Uniform corrosion
of cavities of variable depth that give a rough appearance to the uniformly corroded
surfaces. However, if the anodic region remains in a fixed place, simply growing in
diameter and depth, the mechanism is called “localized corrosion.” This is illus-
trated in Fig. 5.3.
The kinetics of the corrosion reaction is determined by an empiric equation known
as the “Tafel equation.” In this equation, the half-cell over potential (η) is the dif-
ference of the oxidation potential with respect to a hydrogen half-cell, the potential of
which is arbitrarily zero. The Tafel equation for the oxidation reaction is
102 5 Environmentally-Assisted Fracture
iox
gox ¼ box ln ;
i0
where box and i0 are experimental constants. The Tafel equation for the reduction
reaction is
iRed
gRed ¼ bRed ln ;
i0
where R is the ideal gas constant, T is the temperature, z is the transferred load, F is
the Faraday constant, and iLim is the boundary current density in the diffusion layer.
When this equation is plotted in the Evans diagram, the curve for the cathodic
reaction bends downwards, as shown in Fig. 5.5.
A very important phenomenon in corrosion occurs when a thin and very resistant
corrosion product film is formed on the metal’s surface. This film isolates the
E X/X - M M N+ + e-
(Oxidation)
E M/M +
X n+ + n e- X
(Reduction)
iCorr
Oxidation
-
Reduction
+
E Transpassive zone
Passive zone
Transition zone
E X -/X Epp
ip icor ipp
log i
surface from the electrolyte, strongly reducing the corrosion rate. Such a phe-
nomenon is known as passivation, and it modifies the Evans diagram as shown in
Fig. 5.6.
The Evans diagram for passive behavior is divided into four regions:
1. Active zone: Corrosion kinetics is controlled by the electric charge exchange.
2. Transition zone: The kinetics begins to reduce as the potential increases over
points Epp and ipp.
104 5 Environmentally-Assisted Fracture
The usual SCC laboratory test consists in determination of the time of rupture as a
function of the applied stress in a specific environment. The test piece is similar to a
smooth tensile test specimen. The results are plotted on a semi-log scale, so as to
obtain a graph like the one shown in Fig. 5.7. At first, it is observed that at low
stresses, the rupture time is very long, and there is a stress under which the rupture
time is practically infinite. This stress is called SCC threshold stress, and its value
depends on the combination of material and environment.
Just like in fatigue, the SCC general condition is that the applied stress is below
the yield strength, so the crack propagates within an elastic environment. This
means that the crack growth behavior can be analyzed according to linear elastic
fracture mechanics. In such a case, the stress level at the crack tip depends on the
stress intensity factor K. When the crack growth rate (da/dt) versus K data is plotted
on a log-log scale, a graph similar to the Paris graph is obtained, as shown in
Fig. 5.8. In this plot, three clearly defined stages can be observed.
Stress
σ th
log (Time)
5.3 Stress Corrosion Cracking 105
Fig. 5.8 Crack growth rate by ACE, according to the stress intensity factor
Stage I: This shows a K threshold value below which the crack does not
propagate. It strongly depends on the microstructure, and brittle fracture dominates.
The crack growth rate can be estimated by an equation similar to the Paris Law.
Stage II: The crack growth rate shows little or no dependence on K. It is said
that the crack growth rate is then controlled by corrosion reaction kinetics, since the
crack grows by anodic dissolution of the crack tip, and therefore the crack growth
rate depends on the corrosion rate (icorr).
Stage III: The crack growth rate is very sensitive to the stress level. Static modes
of fracture are observed. The process ends when K reaches the fracture toughness
value in the testing environment.
The cracking mechanism, and consequently the characteristics of the fractured
surface, varies from one stage to the next, as shown in Table 5.1.
The effect of the microstructure in SCC is the most complex of all fracture
mechanisms. The highly localized cracking process makes it very sensitive to local
SCC cracks generally initiate at corrosion pits (unless there is a preexisting crack or
a sharp stress concentrator), as shown in Fig. 5.9. Usually, the nucleated cracks are
very short, but a stable crack growth mechanism is quickly established, making
Stage I very brief.
SCC fractured pieces have a brittle appearance, and there is clear evidence of
corrosive attack on the fractured surface. The crack path is tortuous, which makes
Fig. 5.10 Example of an SCC crack and its fractured surface at the macroscopic scale
the fractured surface very rough, as in the example given in Fig. 5.10. The external
surface of the piece can also be corroded, although not as severely as the fractured
surface, unless the piece was collected a certain time after the failure and it had an
intemperism attack. In this last case, the fractured surface is completely covered by
a thick layer of corrosion products, as shown in Fig. 5.11.
At the macroscopic scale, SCC fractures frequently show branches and sec-
ondary cracks, which make the fractures very rough, with plenty of ratchet marks,
steps and facets similar to a cleavage fracture. They can also have pits, ridges and
marks that look like fatigue striations, as seen in Fig. 5.12. The tortuous path of
SCC is the result of the fact that the crack follows the most energetically favorable
path, which is quite variable through the microstructure.
108 5 Environmentally-Assisted Fracture
Fig. 5.12 Example of a tortuous and branched SCC crack making a very rough fractured surface
through the anodic type grains, which are dissolved, making the crack TG. While in
the passive region, the grain boundaries have more energy, so the passive film is not
formed or it is easily broken, turning the grain boundaries themselves into anodes,
thus favoring an IG fracture.
110 5 Environmentally-Assisted Fracture
Intergranular (IG)
E
Transgranular (TG)
Log i
Free surface
External Adsorption
environment
(Temp., pH, Film
Deposit layer Anodic
pressure, formation Process
formation dissolution
chemicals) and rupture zone
M
A- M M Embrittlement
Internal and plastic
Transport of oxidant environment strain
A-
A
-
A-
A M
M
Fractured surface Diffusion of
chemical spices
Crack
Extension by Extension by
dissolution brittle fracture
STAGES
Strain I II III
ε
(dε/dt)
Decreasing (d ε/dt)
(d ε/dt) Increasing
Constant
εO
Time
de=dt ¼ Crn ;
where C is a constant and n is the creep exponent. According to this law, the greater
the stress, the greater the strain rate. In secondary creep, the constant strain rate
mechanism is a thermally activated process, and therefore the strain rate can be
represented by an Arrhenius type equation:
de
¼ Aeð RT Þ ;
DH
dt
where DH is the activation energy, T is the absolute temperature, R is the ideal gas
constant and A is a constant that depends on the material. This equation predicts that
higher temperatures exponentially increase the strain rate, thus it can be stated that
temperature has a much higher effect than the stress in the creep rupture time.
At the macroscopic scale, creep fractures present mild plastic deformation, little
neck strain, multiple cracking and surface roughness. The intense intergranular
cracking in creep frequently causes the fracture to occur with little plastic defor-
mation, as shown in Fig. 5.19. This may be mistaken for creep fractures with brittle
or SCC fractures, however, the observation of intergranular cavitation in a metal-
lographic specimen, plus the typical high temperature oxidation in the fractured
piece, should help in clearly identifying a creep failure.
As has been mentioned, creep fracture is typically intergranular. The basic
mechanism is nucleation, growth and interconnection of grain boundary voids. The
formation of grain boundary voids may be due to three mechanisms:
1. Grain boundary sliding.
2. Vacancy condensation at grain boundaries.
3. Creep deformation.
The grain boundary sliding is a consequence of the loss of mechanical strength
experienced by grain boundaries at high temperatures. The grain boundaries ori-
ented favorably to the local maximum shear stress will slide, causing decohesion at
the triple joints, forming wedge-like voids known as “W” type. Conditions that
favor W type voids are: high temperatures, above 0.6TM, and high stresses. The
Fig. 5.20 schematically shows the formation of W voids.
Vacancy condensation at grain boundaries, on the other hand, leads to the for-
mation of spherical voids along the grain boundaries, called “r” type. Their growth
is controlled by vacancy diffusion and stress, according to the following equation:
dr=dt ¼ C DV r m rn ;
where dr/dt is the r void growth rate, Dv is the vacancy diffusivity, r is the void
size, n is the power law creep exponent and m is an experimental constant. The
formation of type r voids is favored by relatively low temperatures and low stresses.
Type r cavities are preferably formed in grain boundaries, closely perpendicular to
5.4 Creep Fracture 115
Grain Tripe
boundary joint
slip Wedge-
like void
Type r voids
the direction of the maximum tensile stress, as shown in Fig. 5.21. The presence of
precipitates on the grain boundaries restrains the grain slip, and thus type W cavity
formation, but has no effect on the formation of type r voids. Figure 5.22 shows
grain boundary cavitation in a piece of steel with creep failure.
Creep fractured surfaces are typically intergranular and show cavities on the
grain boundary surfaces, as shown in the example in Fig. 5.23.
Fig. 5.22 Grain boundary cavities in a piece of steel with creep failure
This makes creep failure possible, even when the applied stress is not sufficient to
cause overall creep deformation and rupture. In this case, high stresses and strain
rate in the process zone accelerate grain boundary cavitation, as illustrated
schematically in Fig. 5.24.
Creep crack growth rate (da/dt), according to the mechanism grain boundary
cavitation and sliding, is controlled by the crack tip stress, which depends on
5.4 Creep Fracture 117
Void growth
Process zone at higher
than nominal stress
Crack extension
Fig. 5.24 Schematic representation of creep crack growth by grain boundary cavitation and
sliding
K. Experimentally, it has been found that creep crack growth rate, when the applied
stresses are below the yield strength at the service temperature, follows the equation
da=dt ¼ C ðK Þn :
On the other hand, when crack growth is controlled by the creep diffusive strain,
the growth rate can be estimated by the C* parameter (defined by dJ/dt, where J is
the J integral). Logically, both grain boundary cavitation and diffusive strain crack
growth mechanisms may appear simultaneously in a creep failure, so the super-
position principle can be applied to estimate the creep crack growth rate as
ðda=dtÞtot ¼ C1 ðK Þn þ C2 ðCÞm :
by the conversion from dissolved hydrogen to gaseous hydrogen in the trapping site
creates a stress concentration at the edge of the trapping site, cracking the metal and
forming a larger cavity filled with molecular hydrogen. In this process, the metal is
embrittled by the dissolved hydrogen, facilitating the propagation of the crack. This
mechanism, known as a “pressure mechanism,” is schematically represented in
Fig. 5.25.
If the crack is formed near the half thickness, HIC will form an arrangement of
planar cracks that propagate parallel to the plate wall. If the crack is formed near a
free surface, the internal pressure may cause the plastic deformation of the
remaining wall thickness, forming a blister. Figure 5.26 shows the macroscopic
aspect of a blister and a bulged HIC planar crack (known as lamination) formed in a
sour oil pipeline.
Although macroscopically different, blisters and laminations are produced by the
pressure mechanism described above, so the fractured surfaces of blisters and
laminations are similar. In both cases, the hydrogen-induced cracking generates
multiple cracks in planes at different depths across the plate thickness, forming a
stepped crack array, as shown in Fig. 5.27.
At the microscopic level, HIC fractured surfaces have a brittle appearance,
mainly because the dominant cracking mechanism by decohesion of the ferrite and
perlite bands interfaces, as shown in Fig. 5.28.
Also, it is common for the HIC crack tips to bend towards the neighboring
cracks, because of the interaction of the approaching crack tip stress fields.
Likewise, it is common for the material between the approaching cracks to show
some plastic strain, because of the great stress concentration at these points, as
shown in Fig. 5.29.
5.5 Hydrogen-Induced Cracking 119
Fig. 5.26 Macroscopic aspect of a blister (above) and a lamination (below), formed by
hydrogen-induced cracking in pipeline steel
Fig. 5.27 Step-wise hydrogen-induced cracks in pipeline steel. Polished surface, no attack
Fig. 5.29 Microscopic aspect of HIC in a steel exposed to a sour environment. Scanning electron
microscope, attacked with Nital 3
5.5 Hydrogen-Induced Cracking 121
Fig. 5.30 Nonmetallic inclusion that was the nucleus of an HIC fracture
Due to the constant presence of sour hydrocarbons in the oil and gas industry, HIC
has been extensively studied since the 1980s, although most research has focused
on the prevention of the phenomenon, that is, the design of HIC-resistant steels,
resulting in the so-called “sour service grade steels.” In these steels, Sulphur content
is reduced in order to minimize the content of MnS inclusions; in addition, the steel
is treated with calcium or rare earth elements to control the shape of inclusions,
making them spherical, because it was found that this shape is less effective as a
nucleus of HIC than elongated MnS inclusions.
On the other hand, the kinetics and fracture mechanics of HIC have received less
attention from both steel manufacturers and researchers, so nowadays, the recom-
mended practices and fitness-for-service standards simply evaluate HIC as a local
metal loss, assuming that the thickness section that shows step cracking does not
carry any stress. Such assumptions are often overly conservative, and allow for the
presence of small-size HIC cracks (usually a few inches long), so that many
components that contain HIC cracks end up being replaced or repaired, at very high
costs.
The main characteristics of the kinetics of HIC can be summarized as follows:
Fig. 5.32 shows the straight beam ultrasonic mapping contours of HIC in steel
plates. The HIC was artificially induced by cathodic charging, which reproduces
Fig. 5.32 Ultrasonic mapping of HIC nucleation and growth in a steel plate exposed to cathodic
charging (current density of 2.48 mA/cm2). Left, 48 h of exposure; right, 840 h of exposure
5.5 Hydrogen-Induced Cracking 123
Fig. 5.33 HIC kinetics measured from ultrasonic mapping of cathodic charged steel plates
HIC in exactly the same way as it occurs in real life sour service. It is observed that
most cracks initially nucleate almost simultaneously in a few days. These cracks
grow individually at first, while a few new cracks continue appearing. Eventually,
the individual cracks begin to interconnect with each other, forming larger cracks.
When most of the cracks have interconnected, the crack growth stops almost
completely, with the larger cracks growing only by the interconnection of new
cracks.
The kinetics of HIC can be determined from the mappings described above,
either by average crack length or cracked area. The observed HIC growth rates are
as shown in Fig. 5.33. It can be observed that the kinetics follows an exponential
law in the early moments of growth, when most of the cracks grow individually,
shifting to a very slow growth rate when most of the cracks have interconnected
with each other; after some time, the growth almost stops. It is worth mentioning
that it is not yet clear whether the cracking stops because the entire exposed area of
the test plate has been cracked, or because after reaching a certain size, the HIC
cracks do not get enough hydrogen influx to continue growing (or may be hydrogen
leaks). This behavior has been observed in the pipeline industry, where steel pipes
exposed to sour environments for several years contain large HIC areas formed by
many interconnected cracks, which nonetheless stopped growing after some time.
Chapter 6
Failure Analysis of Fractured
Components
Failure analysis texts present several definitions of failure, but generally, a failure
can be defined as “the condition that makes a component inoperable, work
improperly, be inefficient, or that produces a risk.” According to this definition,
the failure of an engineering component such as a piece of machinery or a structural
component do not necessarily imply excessive stresses, material deterioration or a
fracture; nevertheless, within the context of this book, the only failures considered
are those that are a consequence of the formation and propagation of a crack, so the
failed component shows a fracture.
Failures that involve fractures are referred to as mechanical failures.
A mechanical failure is the one that occurs by applying loads or stresses or by
strains in the material, up to a level that the material is unable to resist, which results
in overstraining or fracturing of the component, rendering it useless. It could also
happen that stresses and strains are within the permissible limits and the failure
takes place because of deterioration of the material; thus, it is said that it is a
material failure. In the industry, the definition of failure is broad and has to do with
an unexpected condition that makes a component unable to continue in service;
thus, it is said that the component has failed in service.
According to the classic sequence of mechanical behavior—loads, stresses,
elastic strains, plastic strains and fracture—a material failure indicates a plastic
strain condition. For example, a wheel that has lost its circular shape due to plastic
strain will not work properly, and so it has failed. The assumption that a failure is
associated with plastic strain and that a fracture comes after that encourages
designers to keep the applied stresses below the material’s yield strength, hoping
that this will guarantee that plastic strain will be prevented, and so will fracture.
Considering the latter, in-service mechanical failures might only occur as the
result of overloads, that is to say, when the maximum allowable loads are exceeded.
However, despite such design prevision, it is common for fractures to occur at
stresses far below the yield strength, either because there are cracks or stress
concentrators, and this is why mechanical failure may occur, even when the
component is operating within the design limits.
According to the above, a failure that involves a cracking or fracture may occur
due to any one, or any combination, of the three following causes:
I. The allowable stresses are exceeded, either by accident, negligence, abuse
or conditions that added unexpected loads, like winds or earthquakes.
II. The material has insufficient mechanical strength, either as a result of
manufacturing defects, incorrect specifications or because mechanical
properties deteriorated during service.
III. There are defects that reduce the component’s capability to bear loads.
These defects may be cracks, stress concentrators, thickness loss (local or
general), holes, grooves, gouges, dents, etc.
By knowing the mechanical behavior of materials, fracture mechanics, damage
mechanisms, engineering design and manufacture processes, the failure analysis
team will be able to identify, without ambiguities and in a scientific way, which of
these causes was the main contributor to the failure, and the team will issue rec-
ommendations for the prevention of similar failures in the future.
It is important to bear in mind that failure analysis is not limited to failures
occurring in service, since failures can occur during manufacturing or be inten-
tionally induced in a controlled experiment. Whatever the case, the main objective
of failure analysis is to determine the causes of failure, establish the relations with
intrinsic and extrinsic factors and to give recommendations to prevent similar
failures.
6.2 Procedure for the Analysis of Mechanical Failures 127
complemented with animations, simulations or scale models, and the failure analyst
must be skilled in both the writing of technical reports and oral communication.
Based on the above, the best practice for a failure analysis is that it should be
carried out following a systematic procedure, in which the material testing tech-
niques, non-destructive tests and analytical procedures are applied in a scientific
way to learn the true causes of failure. This is particularly relevant when the failure
analysis is part of an expert witness to help a judge resolve a controversy. The role
of an expert witness is to provide the technical and scientific knowledge the judge
lacks, therefore the failure analyst should have a knowledge superior to that of an
average person, so he or she can provide convincing proof of what caused the
failure, in congruence with scientific and technical knowledge about the analyzed
failure. Frequently, the expert witness will argue against other expert’s investiga-
tions, whose conclusions may be different or contradictory. That is why failure
analysis must follow the scientific method and his/her report must present clear and
straightforward conclusions regarding the cause of failure; explaining the premises,
reasonings and foundations pertaining to the science or technique upon which the
conclusions are based. This way, it will be easy for the judge to assess which
investigation is the correct one.
Just as it is important to carry out failure analyses following the scientific
method, it is equally so to know and acknowledge judgment mistakes that may
occur during the execution of a failure analysis. Even if mistakes are made inad-
vertently, they can invalidate a study, or even worse, they may lead to erroneous
decisions that will cause other failures in the future. The main causes of mistakes in
judgment in the failure analysis process are:
• Making value judgments. This means considering a conclusion as valid just
because it comes from a source of high ranking authority, without having evi-
dence or arguments that support it. This error includes failure analyses based on
reports made by experts of high prestige who have analyzed similar but not
identical cases, or analyses made by simple comparison with cases reported in
handbooks or manuals. The sources that lead to value judgments can be, but are
not limited to: popular beliefs, opinions from the majority, persons with a
graduate degree who are nonetheless not experts in the field, newspapers,
magazines, social networks and other mass media that may have a general
knowledge of the failure conditions or circumstances, but lack scientific validity.
Also, it is a value judgment to consider a failure analysis valid just because it
was made by a renowned company or institution, even though the persons
conducting the study do not have sound knowledge or did not follow a scientific
methodology.
• Relying on intuition and common sense. Intuition is a mode of thinking that
reaches conclusions without logical reasoning; it is thus “hunches” that give
validity to certain hypotheses, purely through the power of the individual
choosing to believe in them. Common sense, on the other hand, is a knowledge
based on experience or simple reasoning, but without actual knowledge of the
facts or phenomena that caused the failure. The problem with intuition and
130 6 Failure Analysis of Fractured Components
on-site
examination
Collection of Analysis of
samples and related
shipping components
Macroscopic
examination
Applied stress,
failure load and
critical defect size
calculations
Analysis of
results
Sample
preservation
Report
be known, as well as whether there was any presence of harmful species in the
feed, the volumetric or mass flow, as well as the physical and chemical prop-
erties of the processed products. It is also necessary to know whether there was
any extraordinary operating condition before and/or during the failure and if
there were any other abnormal events that could be related to the failure.
It is necessary as well to determine whether it was the first failure of the com-
ponent or if there have been similar ones. A copy of the previous failure analysis
reports must be obtained (if there are any). Collect any videos, photos or audio
recordings of the moment the failure occurred. A design drawing of the com-
ponent must be requested or made, along with the plot-plant blueprints. Make a
draft of the place where the failure happened.
systems (if there are any) should be collected, and it should be learned whether
or not they were operational during the failure. Damage to other components
must be localized, as well as materials, substances and fragments that were
spilled, expelled or thrown out of the failure location, and finally, the samples to
be taken to the laboratory must be selected and collected, with appropriate
specifications as to their identification, packing and delivery condition. At this
point, the analyst’s judgment is important, because faulty selection of pieces of
evidence may lead to mistaken conclusions. Sample gathering must be carried
out according to the recommendations of step 6 of this procedure.
5. Photographic record. Photographs shall be taken as necessary in order to have
a record of the condition of the failed component, as well as the equipment or
facility where it was installed and any other details about the failure, even if they
may initially seem unimportant. Photographs must include whole failed piece(s),
and show the fractured surface, damaged areas, areas of local thinning and mass
loss, worn zones, dents and gouges, corrosion, etc. If possible, photographs
showing how the component looked when installed should be taken, as well as
photos of similar components that did not fail. Reference scales such as rulers or
objects of a known size should always be used to give a clearer idea of the
component’s size. It is recommended that the photographs be reviewed to check
their quality before leaving the failure site. The failure analysis staff taking
photographs must be aware of the risk conditions that may be encountered
during this activity, above all when the inspection is conducted at sites with
explosive atmospheres, spilled harmful substances, sharp edges, very hot or
burning materials, the potential for objects falling down, and so forth.
6. Evidence and sample collection. The selection of any pieces, parts, fragments
and/or residue associated with the failure that are susceptible for analysis shall
be made. Whenever possible, it recommended that components similar to the
failed component be collected, even if they have not failed, in order to make
comparisons. Collection of samples must meet the requirements of identifica-
tion, packing and transfer conditions for their proper preservation.
For this activity, it is recommended to follow the ASTM E1188 “Practice for
Collection and Preservation of Information and Physical Items by a Technical
Investigator,” which, in summary, requires that the collected samples be handled
in such a way that they are not subject to alterations, damage or contamination.
Special care must be taken with packing the containers used for shipping evi-
dence. The identification tag of the collected evidence must include, at a mini-
mum: incident date, collection date, part name and ID code (if there is any),
incident investigated, name and description of collector and shipping address.
7. Preparation and cleaning of samples for fractographic examination. This
activity includes both preservation of the fractured piece and preparation of the
specimen for laboratory examinations. It is important that preservation and
cleaning are done properly, because pieces may suffer physical or chemical
damage, and thus relevant features of the failure and its mechanisms may be
134 6 Failure Analysis of Fractured Components
damage mechanisms associated with the mechanism that caused the failure. It is
important to point out that the microscopic examination should not be done
before having completed the macroscopic examination, because there is a risk
that the microscope will reveal small zones that are not representative of the
general fracture process, leading to faulty conclusions. The microscopic
examination must be carried out from low to high magnification; it is advisable
to use macro-photography or a sketch of the fracture as a guide to quickly
identify the desired zones to observe. Unless it is necessary, taking too many
pictures must be avoided, as well as making mosaics of a large number of
photographs to show large areas, because they make it difficult to edit the
written report.
The microscopic examination can be carried out using optical stereo micro-
scopes or scanning electron microscopes, although it need not be limited to these
techniques. The stereo microscope is used for quick observations at low mag-
nification (less than 100) in samples with minimum preparation, but it is
limited by its low magnification power. The SEM has several advantages, such
as: high resolution (up to a few nanometers), tri-dimensional appearance of the
images and sensitivity to chemical and electromagnetic properties of the mate-
rial, meaning that the contrast, in addition to the topography, is related to the
elemental composition, electrical conductivity and density. Furthermore, every
commercial SEM can be equipped with X-ray analysis devices, which allow us
to carry out qualitative and quantitative chemical analysis in selected areas, fine
particles and thin layers. All of this makes the SEM the most versatile tool for
fracture investigations.
11. Laboratory-tests. Laboratory tests are used for determining the physical and
chemical properties of the fabrication materials, so it can be investigated
whether the material meets the design specifications or if they were altered
during service or by external factors.
Laboratory tests may also be carried out from samples collected at the failure
zone, to determine whether the material was homogeneous or the material
properties in the failure zone changed significantly with respect to the nominal
properties of the component, especially if the failure is highly localized.
Likewise, it is recommended that the mechanical properties of similar compo-
nents that have not failed be evaluated, in order to determine whether the cause
of failure is related to an alteration or low quality of the fabrication material.
The customary laboratory tests are, though not limited to, the following:
i. Metallographic examination.—Observation of the microstructure is an
essential part of a failure investigation, since it is closely related to the
mechanical properties and damage mechanisms. Actually, all engineering
materials can be examined through metallographic techniques, so it should
be a routine part of the failure analysis of fractured components. The pro-
cedures for the metallographic test are described in the ASTM E3, ASTM
136 6 Failure Analysis of Fractured Components
E1122 and ASTM E1382 standards. This test provides the investigator with
information about the microstructure (grain size, second phases, inclusion
content, etc.), fabrication process and heat treatment of the material, and
allows for the detection of microscopic defects like porosity, micro-cracks,
segregations, etc. Other characteristics, like corrosion layer thickness, surface
modifications, intergranular oxidation and crack path, can also be revealed
by metallographic examination.
ii. Hardness testing.—This is the most frequently used mechanical test in
failure analysis. Hardness is the resistance of a solid to being penetrated or
scratched. For metallic materials, it is evaluated according to the ASTM E10
standard. Hardness is a complex property that depends on the elastic and
plastic behaviour of the material, but it is directly related to mechanical
properties such as tensile strength, ductility and wear. Hardness testing is
useful for detecting cold work, heat softening, hard spots, extension of the
heat-affected zone in welds, etc. In general, surface hardness is considered
nondestructive, except when the indentation load produces micro-cracks or
deforms the test piece.
iii. Tension testing.—This is used for determining the material’s yield strength,
ultimate tensile strength elongation and reduction of area. The test is carried
out according to the ASTM E8 standard. It has to be done whenever possible,
because most material specifications are based on the tension properties. It is
useful for determining whether the material met the design specifications or
if it experienced an alteration during service.
iv. Charpy impact testing.—This test is frequently used in failure analysis,
because it helps determine the tendency towards brittle fracture, the
ductile-brittle transition temperature and the material’s fracture toughness. It
is done according to the ASTM E23 standard.
v. Fracture mechanics testing.—These tests include plane strain fracture
toughness (KIC), elasto-plastic fracture toughness (JIC) and critical crack tip
opening displacement (CTODc). The test methods are covered in the ASTM
E1820 standard, while fatigue crack growth rate is covered by the ASTM
E647 standard. Because these tests are much more complex and take longer
than other mechanical property testing, they are seldom done. Fracture
mechanics tests are useful for determining the crack’s critical size and the
remaining strength of the failed component, whereas crack growth rate tests
are useful for determining the time that the crack took to grow from its initial
size to its critical size, thus determining the service life of a fractured
component.
vi. Tests in simulated environments.—In very deep failure analyses, tests for
evaluation of material properties in an environment that simulates the failure
service environment may be required, either to obtain the in-service
mechanical properties or to reproduce the failure under controlled condi-
tions. The most common test of this type is the high temperature tension test.
Laboratory tests in simulated environments provide information about the
mechanical behavior of the material under the actual service conditions,
6.3 Failure Analysis Procedure of Fractured Components 137
whereas failure tests allow for verifying the cause of failure through com-
parison of the test specimen characteristics with the actual failed component.
Special care must be taken with accelerated tests, because they may not be
fully representative of the current material performance or the damage
mechanism that caused the failure.
vii. Chemical analysis.—Chemical analysis can be elemental (when the element
content is determined) or for the identification of chemical species (mole-
cules). It can be qualitative, when it only identifies the chemical species, or
quantitative, when, in addition to identifying the species, the concentration is
measured. Chemical analysis is another test that has to be conducted in a
failure analysis whenever possible. It is useful for the correct identification of
the material and to determine whether the material meets the design chemical
analysis specification. Nowadays, there are portable devices that allow for
quick and nondestructive chemical analysis, but it is highly recommended
that the chemical analysis be performed at a certified laboratory. It is also
important to be aware of the representativeness of the samples taken for
chemical analysis, since most engineering materials may not be completely
homogeneous.
viii. Physical and chemical property testing.—Other material properties, such
as fatigue life, corrosion rate, crack propagation rate, wear rate and creep
rupture time, as well as physical properties like thermal expansion coeffi-
cient, thermal conductivity, density, etc., may be needed to conduct a
complete failure analysis. The tests should be done following standard
procedures, and they must be done under service conditions as close as
possible to those of the failed piece.
12. Simulations, bench tests and in situ testing. When resources assigned to the
failure analysis allow it, computerized simulations, pilot model tests, bench
tests or tests on real components may be used to reproduce failure scenarios
and, with their results, verify a cause-of-failure hypothesis. In the analyses of
failures involving fractures, stress simulations by finite element methods pro-
vide valuable information about the stress state, stress concentrations, strains
and displacements of the component under service or eventual load scenarios.
Additionally, they allow us to conduct a parametric analysis that helps us to
better understand the failure process. Caution is necessary in the selection of
material properties and boundary conditions, as well as determination of a
suitable criterion for limiting states, all of this in order to make the simulation as
close as possible to reality.
It is important to state that a simulation must never be the only support for a
failure analysis conclusion, but should rather serve merely as a complementary
part of the analysis to verify a cause hypothesis. This is because simulations
always consider simplifications, idealizations and assumptions that make them
differ, to a lesser or greater extent, from the actual conditions that led to a failure.
138 6 Failure Analysis of Fractured Components
IV. Methodology.- This section describes the activities carried out for the failure
analysis, including the description of field tests, nondestructive inspections
and laboratory tests that were performed for the study. In the case of tests, it
will be sufficient just to mention the testing standard, the number of tests
done and the materials and equipment used.
V. Testing results.- The testing results should be presented in the usual fashion
for a technical report, complete with tables, graphs, pictures, etc. Any rele-
vant comment or observation about the nature of the results may be included,
but conclusions must not be foretold. When calculations are done, the
equations, formulas or procedures must be described and the references cited.
Results of simulations, bench tests and in situ tests must be described in this
section. Once again, it is recommended to present only relevant results and
confine those not leading to conclusions to an appendix.
VI. Analysis of results.- In addition to describing the reasoning, logical
deductions and inferences that support the determined failure causes, this
section also describes the complementary analyses, such as continuum
mechanics and fracture mechanics calculations, computer simulations and
analysis and interpretation of testing results. It must be remembered that
calculations and simulations are a tool for verifying a failure hypothesis,
therefore they are part of the result analysis, but they are not evidence of the
failure event. The arguments for rejection of other hypotheses are presented
in this section. A brief description of the recommendations for prevention of
similar failures derived from the study may be given at the end of this
section.
VII. Conclusions.- The conclusions must be numbered, and it is recommended
that the first conclusion be the cause of failure, postulated as definitive or
possible. The subsequent conclusions may summarize the results and
describe the reasoning that supports the cause-of-failure conclusion. It is
optional to include conclusions about rejected hypotheses (although not
advisable, because it may cause confusion). The recommendations may be
stated along with the final conclusions or presented in a separate section of
the report.
The format of the written report, such as letter type and size, paper size, margins,
bonding, etc., shall be agreed upon with the customer or simply use the typical
format for technical reports, that is: white paper, paper size letter or A4, letter
size 10, 11 or 12. It is important to verify that the letter type is one in which
upper and lower case letters and numbers are clearly distinguishable. It is rec-
ommended that the pages be printed on one side and the page numbers be in the
X of Y format. The writing style must be according to the standards of technical
documents. In the case of translations into another language, it is recommended
that the translator be a technician, with expertise in the engineering field and
terminology of the failed component and the system in which the failure
occurred.
6.3 Failure Analysis Procedure of Fractured Components 141
Practically, in all engineering designs, it is assumed that a material will fail if the
applied stresses surpass the yield strength, so that the component becomes plasti-
cally deformed. Therefore, determining whether the applied stresses caused plastic
deformation is fundamental to the failure analysis of mechanical or structural
components.
The continuum mechanics theory, from which the concept of stress was intro-
duced, was proposed by Augustin Louis Cauchy in 1820. This theory states that the
stresses within a loaded static body are defined by the stress tensor, which is a
matrix composed of nine vector components: three normal (r) and six shear (s)
stresses, as shown below:
0 1
rxx sxy sxz
r ¼ @ sxy ryy syz A:
sxz syz rzz
The first sub-index indicates the area on which the stress is acting and the second
is the direction of the stress. This with respect to a coordinated set of orthogonal
axis x, y and z.
The Cauchy stress theory predicts that there will always be an orientation at
which the shear stresses are zero and the normal stresses are maximum and mini-
mum. These normal stresses are called Principal Stresses, thus the principal stress
tensor is expressed as
0 1
r1 0 0
r¼@ 0 r2 0 A;
0 0 r3
142 6 Failure Analysis of Fractured Components
where r1 is the maximum principal stress, r2 is the middle and r3 is the smallest
(or the most negative) stress. The theory also establishes that there is always an
orientation at which the shear stress has a maximum value, and this is at 45° from
the direction of the principal stresses. The method for determining the principal
stresses and the maximum shear stress was introduced by Otto Mohr in 1882 and is
known as Mohr’s Circle.
Yield strength is the stress at which a material begins to plastically deform, while
the ultimate tensile strength is the maximum stress that the material can withstand,
before going into plastic collapse failure. However, these definitions only apply for
the uniaxial tension stress state, that is, when the stress tensor has only one com-
ponent of tension stress, for example, rxx and the rest of the components are zero.
To determine whether there is yield under a combined stress state, that is to say,
when there are two or more stress components different from zero in the stress
tensor, a continuum mechanics yield criterion needs to be used. The most widely
used failure criteria are the Tresca and the Von Mises, which will be described next.
The Tresca criterion, was introduced in 1864 by Henri Edouard Tresca, and it is
known as the maximum shear stress criterion. It establishes that plastic strain will
begin when the maximum shear stress surpasses a critical value. For any stress
state, the maximum shear stress is given by the difference between the maximum
and the minimum principal stress, therefore, the Tresca criterion is applied by the
following equation:
r1 r3 ¼ ro :
Richard Von Mises determined in 1913 that yielding begins when the effective
stress value is above a critical value. His criterion is given by the following
equation:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
r0 ¼ pffiffiffi ðr1 r2 Þ2 þ ðr1 r3 Þ2 þ ðr3 r2 Þ2 :
2
The Tresca criterion is the most widely used for the analysis and design of
structures, because its equation is simpler and predicts smaller failure stresses,
which provides an additional safety factor. The von Mises criterion, on the other
hand, is more precise and has the advantage that it does not require calculation of
the principal stresses, because it can be expressed by an equation that uses the
original stress components, as follows:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 2 2
r0 ¼ pffiffiffi rxx ryy þ ðrxx rzz Þ2 þ ryy rzz þ 6 s2xy þ s2xz þ s2yz :
2
If the component has a pre-existing crack, the continuum mechanics theory loses
validity, because the mathematical singularity introduced by the crack tip does not
allow for calculation of the stress components. Instead, fracture mechanics is used
for the calculation of fracture load and the critical crack size of cracked components.
In order to calculate the fracture load, one must start from the intensity stress
factor of linear-elastic fracture mechanics, developed by George R. Irwin in 1954,
whose general form is
pffiffiffiffiffiffi
K ¼ Yr pa;
where K is the stress intensity factor, Y is a geometric factor, r is the applied stress
and a is the crack size. When K is replaced by fracture toughness KIC, the failure
stress rf can be calculated:
KIC
rf ¼ pffiffiffiffiffiffi :
Y pa
P
rN ¼ :
BðW aÞ
If the previous equation is multiplied and divided by BW, and the term P/BW is
replaced by the applied stress rA, it can be stated that
a
rA ¼ rN 1 :
W
By assuming that the plastic collapse failure will occur when rN equals the
ultimate tensile strength ruts, the remaining strength equation by the net section
criterion has the form
144 6 Failure Analysis of Fractured Components
Failure
stress
Design
σf = σuts (1 - a/W)
stress
σf = KIC/Y√πa
Operation
stress
a0 ac
Crack size
Fig. 6.2 Failure assessment diagram, showing the crack size range in which fracture mechanics
(solid line) and net section criteria (dotted line) are valid, as well as the critical crack size and the
maximum tolerable crack size
a
rf ¼ ruts 1 :
W
By plotting this equation in the fracture mechanics FAD, a set of plots similar
those shown in Fig. 6.2 will be obtained. The crack size range for which the
fracture mechanics’ remaining strength is valid is where rf calculated by the net
section criterion is greater than the rf calculated by fracture mechanics. The crack
size that gives a failure stress equal to the operation stress is the critical crack size
(aC), and it will indicate the cut-off of the right-side FAD curve. On the other hand,
if the stress failure equation is solved for the crack size using the design stress, a
maximum tolerable crack size is obtained (a0). The region between the maximum
tolerable crack size and the critical crack size defines a zone in which the com-
ponent may tolerate cracks, but there is still a risk of failure, if the applied stress
goes above the failure; this is the basis for the fitness-for-service assessments. The
shape of the FAD constructed under these considerations is shown in Fig. 6.2.
The value of aC can be calculated by solving the fracture mechanics stress failure
equation, using the operation stress rOP and KIC values, thus
2
KIC
aC ¼ :
ðpYrOP Þ2
In the 1980s, Milne and coworkers submitted the CEGB R6 report, which describes
a method for assessing the remaining strength of a cracked component, under
6.4 Failure Analysis and Fracture Mehanics 145
Applied K Brittle
fracture
FAILURE
Load line
Ductile fracture
Kr = KI / KIC
Assessment Plastic
point collapse
Material SAFE
KIC
Net
Crack section
size stress Sr = σ / σLIM
brittle, elastic-plastic or plastic collapse failure modes, with a single equation. The
method, initially known as the R6 Code, considers the ratio between the acting
intensity stress factor (K), the fracture toughness (Kc) and the ratio between the
applied stress (r) and a limit strength (rLIM), usually taken as the yield strength.
These ratios are expressed as the parameters
Kr ¼ K=Kc;
Sr ¼ r=rLim :
Currently, this method is known as the Two-Parameter, and has been adopted by
virtually all industry standards for the assessment of cracked components in a wide
variety of structures, pipes, pressure vessels and machinery, including welds. The
equations of the Two-Parameter criterion are continuously revised and updated. The
most widely used equation is the Level 2 of the API579/ASME FFS-1 Part 9, which
has the form
Kr ¼ 1 0:14 Sr 2 ½0:3 þ 0:7 exp 0:65 Sr 6 :
applied load, while the length from the origin to the intersection with the curve is
proportional to the failure load. The calculation of the failure load can be easily
solved by a rule of three, whereas the calculation of the critical crack size requires
interaction and is normally done by parametric analysis with computer programs.
The fracture mechanics equations are useful for making calculations of important
parameters of fractured components. For example, if the applied stress and the
critical crack size are known, the material fracture toughness can be estimated, and
it can then be determined if the material had the correct fracture strength for the
service conditions or if the fracture strength was reduced during service. Therefore,
it can be further known whether some form of embrittlement occurred. If the
fracture toughness and the critical crack size are known, the failure stress can be
estimated, and by comparing it to the normal operational stress, an overload event
can be identified (if the failure stress is greater than the operational stress) or it can
be known whether the material strength deteriorated during service (if the failure
stress is below the operational stress).
The equation for determining fracture toughness from the critical crack size is
pffiffiffiffiffiffiffiffi
KIC ¼ Yr paC :
aC
aC
aC
Fig. 6.4 Identification and measurement of the critical crack size in different fractured surfaces
½ rp
½ rp
Fig. 6.5 Plastic zone size estimation from the shear lip height in ductile fractures
148 6 Failure Analysis of Fractured Components
Fig. 6.6 View of the cylinder crown with four broken bolts. The bolt indicated with the arrow has
a fatigue fracture, while the remaining three show tensile overload fractures
recommendations. The engine failed just a little before an overhauling service that
was scheduled at 30,000 service hours.
Visual examination.- The failure was localized in one of the piston cylinders,
but it caused the rupture of a large number of parts in that zone, such as rods and
bolts. The connecting rod broke the engine block wall. After carefully examining
the fractured fragments, it was observed that in a component called the “cylinder
crown,” there were four fractured fixing bolts. By a macroscopic rough fracto-
graphic examination of the fractured surfaces, it was seen that three bolts showed
overload type fractures, so it was determined that they broke suddenly and in a
single load event, whereas one showed a typical fatigue fracture (Fig. 6.6).
Fractographic examination.- The four bolts on the failed cylinder crown were
collected and taken to the laboratory for a detailed study. Additionally, two other
bolts from a cylinder that did not fail were collected. Figure 6.7 shows the fractured
cylinder bolts. Bolts 1, 2 and 4 had rough fractures, with a slant fracture plane, and
show a neck typical of ductile fractures in uniaxial tension. These characteristics
confirmed that bolts 1, 2 and 4 suffered ductile fractures, by tensile overload. In
contrast, bolt 3 shows a flat fracture with a smooth and shiny aspect, with a sudden
transition to a fibrous and opaque fracture. The beach marks in the flat fracture
confirm that it is a fatigue fracture.
Figure 6.8 shows a microscopic view, taken with an SEM, of the flat fractured
surface of bolt 3, where the typical fatigue striations can be observed.
Analysis of results.- The metallographic examination and chemical composition
of the bolts, including both the failed and the non-failed bolts, meet the specifi-
cations of 4140 T steel, despite the fact that the hardness values of all of the bolts
were within the specification. On the other hand, the nondestructive testing carried
out during maintenance of the engine did not show the presence of cracks or
6.5 Examples of Failure Analysis 149
previous damage, and therefore the idea that the cause of failure was a defective bolt
or an incorrect operation is discarded, leaving the possibility that the fatigued bolt
was under a high stress amplitude, and that that, combined with a reduced fatigue
resistance (due to lower hardness), was the cause of its premature failure, as the
only feasible hypothesis.
The Hardness test results showed that the fatigue failed bolt had a hardness 5%
less than the other bolts (328 Brinell vs.−344 Brinell); this indicated that the fatigue
150 6 Failure Analysis of Fractured Components
Fig. 6.8 Fractured surface of bolt 3, as seen in the SEM, showing fatigue striations
limit (which is about 0.5 UTS) of bolt 3 had an estimated fatigue limit of 76.85 ksi,
which is slightly less than the fatigue limit of the other bolts of 79.4 ksi.
To verify the previous hypothesis, it is necessary to determine the stress
amplitude of bolt 3 and compare it to the material fatigue limit. The first step is to
calculate the stress applied to bolt 3. Linear elastic fracture mechanics predicts that
a cracked body breaks when the mode I stress intensity factor (KI) is equal to or
greater than the plane strain fracture toughness (KIC) of the material. The equation
for determining KI in the bolt can be found in Annex C of the API 579-1/ASME
FFS-1 Fitness for service 2007 standard, which is reproduced below:
C.11.2.1 The Mode I Stress Intensity Factor [19]
pffiffiffiffiffiffi
KI ¼ ðMm rm þ Mb rb Þ pa ðC:233Þ
where,
In the equation C.233, the term rb is zero and rm corresponds to the maximum
tension stress (which is the desired value), therefore, by substituting KIC in
equation C.233, introducing the value of a = 0.22 in. (5.6 mm), measured on the
6.5 Examples of Failure Analysis 151
ac = 0.22 inch
fracture surface of bolt 3, as shown in Fig. 6.9, and solving for rm, the following
equation is obtained:
KIC
rm ¼ pffiffiffiffiffiffi :
Mm pa
The KIC value for SAE 4140 T steel quenched in oil and tempered at 540 °C is
66 MPa√m (60 ksi√in.)1 and Mm is 1.38 (calculated value). Substituting values, the
calculated value of rm in bolt 3 is 52.3 ksi.
The next step is to calculate the nominal stress, meaning the stress introduced by
the specified tightening torque, which is the stress that the failed bolt should have.
The nominal stress depends on the diameter and the thread pass. According to the
information from the engine’s manufacturer, the tightening torque is 40 lb ft and
the bolt diameter is 0.5 in., with a pass of 15 threads/in. In order to calculate the
stress on the bolt, a free web application was used,2 giving a value of ra = 32.3 ksi.
By comparing the latter values, it is observed that the fatigued bolt was under
stresses greater than expected (52.3 ksi) and has a fatigue strength (76.85 ksi) lower
than those that did not fatigue, which were under lower stress (32.3 ksi) and greater
1
Wahab et al. “Effect of Heat Treatment on Fracture Toughness of AISI 4140 Steel”. Minia Journal
of Engineering and Technology, Vol. 33, No. 1, January 2014.
2
http://www.turula.com/cgi-bin/bolt.pl.
152 6 Failure Analysis of Fractured Components
fatigue strength (79.4 ksi). Therefore, it is demonstrated that bolt 3 of the diesel
engine crown was prone to fail prematurely.
Contributing factors.-Since there are no records of prolonged shutdowns of the
diesel engine, nor were repairs made by the operator, the hypothesis that the
operator made any alteration of the cylinder crown bolts is discarded. On the other
hand, it is known that the manufacturer is the only one who makes and sells the
crown bolts, thus the cause of failure is attributed to the manufacturer who installed
bolt 3 with slightly less hardness than the other bolts and applied a greater torque
during installation; both factors combined caused the fatigue life of bolt 3 to be
shorter than the others, thus causing it to fail prematurely.
Failure sequence.- Figure 6.10 shows a drawing of an engine cylinder, where it
can be seen that the bolts attach the crown to the piston and prevent it from falling
into the combustion chamber of the cylinder. Based on this drawing and the pre-
vious analysis, the failure sequence was as follows:
Background.- The pinion gear of a cement mill speed reducer showed teeth
spalling after approximately 21 months of service, as shown in Fig. 6.11. As a
corrective action, a new reducer was installed, but the pinion gear failed again, this
time after only three months of service. Three months later, the whole reducer
failed. All of these failures were considered premature, since the expected life of
these parts is several years.
The function of the reducer is to transmit the motor power to a milling table,
where silica rock is ground as raw material for Portland cement fabrication. The
cement mill operates 20 h a day, 7 days a week, with a load of 100 metric ton/h. By
design, the operation is controlled so as to maintain a constant speed, and the only
stages of transient speed are shut-down and start-up. The reducer is driven by a
154 6 Failure Analysis of Fractured Components
2800 kW electric motor and the reducing stages are: (1) 459 rpm at 2100 kW,
(2) 508 rpm at 2400 kW a (3) 572 rpm at 2600 kW, with output rate of 15.14–
18.45 rpm. The pinion gears and the crown are manufactured of 18 Cr–Ni–Mo
quenched and tempered steel. Figure 6.12 shows a drawing of the reducer, indi-
cating the location of the failed components.
Visual examination.- Figure 6.13 shows several teeth with severe spalling on
the contact surface. The cause of the spalling was identified as a combination of
high contact stresses and surface fatigue. In addition, other damages, like pitting,
polishing and cracking, were observed. All 11 teeth that make up the conical pinion
showed damage.
The pinion gear was in service with a polished finish of a metallic gray color,
and it was free of corrosion, distortions and visible geometric defects. The pinion
shaft was free of distortions and geometric defects, with a polished surface finish
and no indications of corrosion or wear.
Fractographic examination.- Figure 6.14 shows the fractured surface of a
pinion tooth, as viewed in a stereoscope. A smooth fractured surface with beach
marks is observed, which is identified as a fatigue fracture. This surface fracture
was observed in an SEM, but the extensive mechanical damage destroyed most of
the microscopic fracture features; nonetheless, in the undamaged areas, some
ill-defined fatigue striations are visible, as seen in Fig. 6.15. Because several
investigators participated in the failure analysis of the pinion gear, the owner
decided to give only one tooth to each investigator, and no other fractures could be
examined in this particular study.
6.5 Examples of Failure Analysis 155
Fig. 6.12 Drawing of the 100 metric ton/h cement mill reducer, showing the location of the failed
gears
Analysis of results.- The spalling, polishing, pitting and cracks observed on the
teeth of the pinion gear were formed when the gears were subject to high loads.
Since the measured hardness values are within the typical values for this applica-
tion, it is concluded that the tooth damage was caused by an operation load above
the mill design’s capability.
To verify whether the pinion teeth have enough strength to prevent tooth pitting,
the tooth pitting index is used, (K factor), given in Sect. 2.2.3 of the standard API
613 (equivalent to the AGMA9 standard), which is calculated according to the
following equation:
126; 000HP R þ 1
K¼ 2 ;
d ðrpmÞFw R
where HP is the transmitted power in horse power, Fw is the tooth width in inches, d is
the pinion diameter in inches, rpm is the revolutions per minute and R is the number of
crown teeth divided by the number of pinion teeth. In order to make a conservative
calculation, a gear service factor of 1.0 (without safety factor) was assumed. For a 58
Rockwell C hardness of the pinion and crown tooth, K should be less than 440, that is,
if K 400, there is no risk of tooth pitting. Considering the possible minimum
design values for the mill reducer pinion, the following data are obtained:
156 6 Failure Analysis of Fractured Components
Fig. 6.15 Fractured surface of a tooth, taken from a zone identified as Stage II fatigue showing
severe mechanical damage; even so, ill-defined striations can be seen
the motor supplies the needed power to keep the mill table at a steady speed,
independent of the milling load (tons per hour) and the hardness of the material
being ground. The maximum power is limited by motor power, so if the material’s
hardness demands the full 2800 kw power from the motor, it will supply it, but if
the reducer is designed for 2000 kw, it will be overloaded, thus causing the pitting,
fatigue cracking and spalling.
Failure sequence.- According to the specialized literature, pitting, cracks, fati-
gue and spalling of gears occur during service, therefore the failure sequence of the
pinion gear was as follows:
1. Since the beginning of the service, the high cyclic loads induced a mechanism of
surface fatigue on the teeth’s surface below the pitch line, forming pits at least
2 mm in diameter, and both the number and size of the pits increased until a
larger portion of material became detached.
2. The large pits acted as stress concentrators that promoted the formation of
macroscopic fatigue cracks (more than 10 mm long) as the number of cycles
continued to accumulate.
3. The fatigue cracks grew until they reached their critical size and caused the
detachment of larger sections of the teeth, or the fatigue cracks interconnected
with other cracks to cause the fall of a tooth.
4. The teeth and chips fell, obstructing gear movement, thus causing the overall
failure of the mill reducer.
Conclusion.- The failure of the pinion gear was caused by the selection of a
cement mill reducer model with insufficient power capacity (2800 kw), which
induced excessive loads on the gears, thus causing pitting, spalling and fatigue
failures.
failure, the pipeline had a little over 32 years of service and it had been internally
inspected at least two times with MFL tools.
Visual examination.- Figure 6.16 shows the failed pipe section as it was
received in the laboratory. The pipe segment dimensions are: 25.4 cm (10 in.)
diameter, 431 cm (169 in.) length and the measured average wall thickness was
6.54 cm (0.257 in.).
The main feature of the failed pipe is a fish mouth-type fracture, 133 cm long
and with a 5 cm opening, localized along the ERW seam. The fracture consisted of
a single crack, with no branching or secondary cracks. The fracture plane is parallel
to the longitudinal direction, without deflections at the ends. The fractured surface
was covered with atmospheric corrosion products.
No local metal loss was observed on the external and internal pipe surfaces, and
the uniform corrosion metal loss was negligible. The pipe segment was free of
mechanical damage, such as punctures, gouges, dents or scratches.
The expansion of the pipe perimeter in the maximum opening zone was 3.1,
while the pipe perimeter away from the fracture was 86.1 cm. Thus, the pipe plastic
expansion was 3.6%. The thickness in the middle of the fish mouth crack, measured
by ultrasound on a square grid of 5 5 cm, 150 cm long by 30 cm wide, showed a
minimum wall thickness of 6.1 mm (0.240 in.), located at the edge of the crack,
which represents a 2.7% loss of the average thickness of 6.54 mm (0.257 in.).
Figure 6.17 shows a metallographic preparation of the transversal section of the
crack plane, showing that the fracture grew along the ERW seam. It is observed that
the crack path is right along the fusion line of the ERW seam, and the heat-affected
zone adjacent to the fracture plane shows long non-metallic inclusions, aligned
parallel to the ERW fusion line; these non-metallic inclusions go across the full
thickness, as shown in Fig. 6.17.
Fractographic examination.- Samples of the fracture’s center were cut, cleaned
and prepared for observation in the stereoscope and the SEM. Figure 6.18 shows
that the fracture has two zones with different appearance, one, localized in the
middle thickness, is a stripe of approximately 2.0 mm wide, with a granular and
shiny aspect, and other, located at each side of the pipe wall in the form of stripes, is
covered with corrosion products and a series of parallel lines that are preliminarily
identified as beach marks.
Fig. 6.17 Metallographic preparation of the transversal section of the fracture end of the failed
pipe, showing that the crack path was along the fusion line of the ERW seam
O.D
6.54 mm
I.D.
Fig. 6.18 Fractured surface of the center of the “fish mouth” fracture, as viewed in the
stereoscope, showing a shiny and granular fracture in the center and fatigue-like fractures on both
sides of the pipe wall
Figure 6.19 shows the microscopic appearance of the fractured surface of the
crack growing from the ID to the mid-thickness zone at the center of the fish mouth
crack. Despite the corrosion products covering most of the surface, poorly defined
striations can be observed, which allow us to identify the fracture mechanism in that
zone as corrosion-fatigue. The appearance of the fractured surface in the
mid-thickness zone (the shiny stripe), as seen in the SEM, is shown in Fig. 6.20; it
is identified as a cleavage brittle fracture.
6.5 Examples of Failure Analysis 161
20 μm
Fig. 6.19 Fractured surface close to the external edge of the central zone of the fish mouth,
identified as corrosion-fatigue fracture
10 μm
Fig. 6.20 Fractured surface in the middle of the thickness of the central zone of the fish mouth,
identified as a cleave fracture
Fig. 6.21 EDS spectrum of the non-metallic inclusions found in the fracture surface, identified as
MnS
6.5 Examples of Failure Analysis 163
1 2t
Pf ¼ ðYS þ UTSÞ ;
2 D
where YS is the yield strength, UTS is the ultimate tension strength, D is the
diameter and t is the thickness. Taking the laboratory-determined tension properties
and the average measured thickness, the burst failure pressure is
1 2 0:25700
Pf ¼ ð70; 589 þ 81; 223Þpsi ¼ 3901 psi
2 10700
This value is equal to 274 kg/cm2, which is far above the pressure of the pipeline
at the failure location, which was 64.5 kg/cm2. Even more, it is above the design
and hydrostatic testing pressure (1.25 times the design pressure), therefore it is
concluded that the pipe must have had a defect that reduced the rupture pressure
down to a value equal to the operating pressure; such a defect should be the cause of
failure.
The size of the crack-like flaw that may have caused failure at the operating
pressure of 63.5 kg/cm2 and the laboratory-determined mechanical properties of the
10-in. diameter and 0.257-in. thickness pipe was calculated by the method
described in Part 9 of API 579/ASME FFS-1 2007 for a Level 2 assessment,
assuming a semielliptical surface-breaking longitudinal crack, located in the weld.
It is important to mention that the critical crack size was determined by iteration,
because the assessment equation cannot be solved for the crack size. The result
gives a crack 30.48 cm (12 in.) long by 3.25 mm (0.128 in.) wide. It is almost
impossible that a crack-like flaw of this size was a manufacturing defect, therefore
the flaw that caused the burst rupture must have grown in service.
In the fractographic examination of the center of the fish mouth, a fatigue
fracture area that covered approximately 70% of the thickness on both the ID and
OD sides of the pipe wall was observed. It is well known that fatigue is a pro-
gressive form of cracking, therefore fatigue cracks growing on both sides of the
pipe wall may easily exceed the critical crack size, and thus it can be concluded that
a fatigue mechanism induced by the frequent operation interruptions of the pipeline,
to address illegal taps, was the cause of failure.
Contributing factors.- In the metallographic examination of the failure zone,
large non-metallic inclusions aligned parallel to the fusion of the seam were found.
These defects are stress concentrators, because they have sharp tip radii, so they can
serve as initiation sites for fatigue.
Additionally, the fracture surface of the ligament, located in the central strip of
the thickness, showed a cleavage fracture, which indicates that the fusion zone of
the seam weld had low toughness values, perhaps due to an improperly controlled
ERW process.
Failure sequence.- Based on the aforementioned cause of failure analysis, the
sequence of events that led to fracture of the gasoline pipeline can be established as
follows:
164 6 Failure Analysis of Fractured Components
1. During fabrication of the failed pipe, long nonmetallic inclusions were formed
along the line of fusion of the ERW seam pipe.
2. The frequent pressurization and depressurization cycles introduced by inter-
ruptions in operation induced a fatigue crack growth mechanism in the ERW
fusion line, which was favored by a low toughness in this location.
3. The fatigue crack grew until it reached a critical size and the pipe failed by
internal pressure, forming a fish mouth type of rupture.
Conclusions.- The cause of failure of the 10-in. diameter gasoline pipeline was
two fatigue cracks, one growing on the external pipe wall side and the other one
growing on the internal side. The fatigue was induced by the frequent interruptions
in operation involving pressurization and depressurization cycles that were the
driving force for fatigue. The fatigue crack growth and the final failure were favored
by the low toughness of the ERW fusion zone, perhaps due to a poorly controlled
welding process.
References
T.L. Anderson, Fracture Mechanics: Fundamentals and Applications, 2nd edn. (CRC Press,
Boca Raton, 1995)
ASM Handbook, Volume 12: Fractography (ASM International, Materials Park, OH, 1987)
ASM Handbook, Volume 11: Failure Analysis and Prevention (ASM International, Materials
Park, OH, 2002)
ASM Handbook, Volume 19: Fatigue and Fracture (ASM International, Materials Park, OH,
1996)
J.L. González, Mecánica de Fractura Bases y Aplicaciones. Colección Textos Politécnicos, Serie
Tecnologías Mecánicas (Editorial Limusa Noriega S.A., México D. F., 1998)
J.L. González, Mecánica de Fractura (Editorial Limusa S.A. de C.V. Grupo Noriega Editores,
México D. F., 2003a)
J.L. González, Metalurgia Mecánica, Segunda Reimpresión (Editorial Limusa S.A. de C.V. Grupo
Noriega Editores, México D. F., 2003b)
R.W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials (Wiley, New
Jersey, 1976)
A.F. Liu, Mechanics and Mechanisms of Fracture: An Introduction (ASM International, Materials
Park, OH, 2005)
A.J. McEvily, Metals Failures: Mechanisms, Analysis and Prevention (Willey, EUA, New Jersey,
2002)
Metals Handbook, Volume 9: Fractography and Atlas of Fractographs, 8th edn.
(ASM International, Materials Park, OH, 1974)
N.W. Sachs, Practical Plant Failure Analysis (CRC Press, Taylor & Francis Group, Boca Raton,
FL, 2007)