EPRI - Fossil Power Plant Components Failure Analysis Guideline

Fossil Power Plant Components
Failure Analysis Guideline
1012744
Fossil Power Plant Components
Failure Analysis Guideline
1012744
Technical Update, March 2007
EPRI Project Manager

D. Gandy
EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • askepri@epri.com • www.epri.com
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER

(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
ORGANIZATION(S) THAT PREPARED THIS DOCUMENT
B. W. Roberts Engineering Consulting
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail askepri@epri.com.
Electric Power Research Institute, EPRI, and TOGETHER•SHAPING THE FUTURE OF ELECTRICITY
are registered service marks of the Electric Power Research Institute, Inc.
Copyright © 2007 Electric Power Research Institute, Inc. All rights reserved.
CITATIONS
This report was prepared by
B. W. Roberts Engineering Consulting

4715 Rocky River Road
Chattanooga, TN 37416
Principal Investigator
B. Roberts
This document describes research sponsored by the Electric Power Research Institute (EPRI).
This publication is a corporate document that should be cited in the literature in the following
manner:
Fossil Power Plant Components Failure Analysis Guideline. EPRI, Palo Alto, CA: 2007.
1012744.
v
PRODUCT DESCRIPTION
The goal of engineering design is to obviate failures. But this goal is only partially achievable
because of the balance between cost and risk, potential deterioration during service, and the
departure of actual operation from design assumptions. Thus, the utility engineer is periodically
faced with failures which span the gamut of power plant equipment and economic/safety
consequences. Reaching a proper conclusion about the failure mechanism and the associated
root cause is central to the post-failure decisions regarding permanent repair, temporary repair,
replace in-kind, replace with an upgrade, or retire the associated equipment. The goal of this
report is to provide guidance to the utility team that performs the failure analysis with a focus on
metallurgical and mechanical aspects of the failure. This failure analysis guideline provides a
roadmap of the sequential steps for the investigation, the laboratory equipment to be utilized,
typical materials for power plant components, and characteristic damage for specific
components.
Results and Findings

A wealth of generic information already exists regarding metallurgical failure analysis.
However, the existing information is not specifically tailored for the utility engineer and has
redundant and superfluous information which is destined to confuse the non-specialist. Due to
the universal challenges facing utilities in retaining technical expertise in metallurgy, welding,
and mechanical design, the need for a readily-accessible failure analysis guideline was identified
as an EPRI priority. This failure analysis guideline was developed in response to the utility need.
This document has assembled the relevant information in a concise manner and logical order
which fulfils the needs of utility engineers to correctly identify failure mechanism and associated
root cause. Whether the failure investigation is performed in-house or is contracted to other, the
responsible utility engineer can use this failure analysis guideline to be confident in directing the
effort. By following this document, the utility engineer will be able to successfully complete the
failure analysis and will be equipped to make enlightened decisions during the post-failure
period.
Challenges and Objectives

The challenge facing all fossil utilities is to maintain safe and reliable operation while facing a
shrinking pool of expertise in the critical technologies. With retirement of fossil plants
becoming a rarity, the aging equipment will encounter increasing failure rates unless a concerted
effort is made to understand critical failures and react to them appropriately. The understanding
of failures usually involves knowledge spread among several engineering disciples, and the
appropriate linkages and integration must be established in a comprehensive solution. The
information provided in this failure analysis guideline will be valuable to mechanical,
metallurgical, and system engineers at both the power plant and central office in establishing the
vii
root cause for a particular failure and in choosing the best options for repair or replacement to
either minimize repeat failures or to give extended life.
Applications, Values, and Use

Failures in fossil power plant components are common occurrences. The tendency is to treat
failures as single events, to treat the “disease” with immediate actions, and to move on to other
challenges. This attitude may squander the opportunity to establish the fundamental failure
mechanism and the associated root cause. Without gaining such fundamental knowledge, history
is destined to repeat itself as evidenced by repeat failures in the same or similar units. By
following this failure analysis guideline, an understanding of each failure will be documented
and lead to enlightened decisions and greater future equipment reliability.
EPRI Perspective
If it is true that knowledge is power, then lack of knowledge surely equates to weakness. This
failure analysis guideline is one of a series of EPRI documents that seek to gather essential
knowledge from disparate sources and distill it into utility-focused documents to broaden the
knowledge base of utility personnel in an easy-to-comprehend presentation. Most of the
information contained in this failure analysis guideline has been long-established, but it is
scattered in dozens of books and publications, usually only familiar to technical specialists.
While the subject area demands a metallurgical emphasis, engineers from other disciplines will
grasp most of the information and be able to put it to practical use.
Approach
The major objective of this project is to provide a comprehensive and detailed document to
provide guidance to fossil utilities to either perform their own metallurgical failure analyses for a
variety of power plant components or to direct organizations that have been given the
responsibility for performing such work. The information will be valuable to mechanical,
root cause for a particular failure and choosing the best options for repair or replacement to either
minimize repeat failures or to give extended life.
To achieve the goals a comprehensive multi-chapter outline was developed and refined through
successive iterations prior to initiation of the project. Accordingly, the first four Chapters
identify the instruments, equipment, and procedures which are utilized to perform failure
analyses as well as the sequential steps to be followed. The remaining Chapters identify typical
materials of construction, generic damage mechanisms, and characteristic damage for an
assortment of power plant components.
Keywords
Failure analysis
Failure mechanisms
Metallurgy
Metallographic examination
Optical light microscopy
Scanning electron microscopy
viii
ABSTRACT
Failures in components of fossil plant boilers, piping, and turbines are common occurrences and
each encounter raises the issue of whether a “failure analysis” should be performed and, if so,
what are the essential elements of that failure analysis. A comprehensive failure analysis has
three major facets that, when combined, provides a consistent framework to claim a “total
understanding.” These three aspects are: (a) stress and fracture mechanics analysis based on
both design and actual operation, (b) confirmation of actual operation relative to the design
assumptions and identification of departures that are relevant to the failure, and (c) the
mechanical and environmental behavior of materials under both the design scenario and actual-
operations scenario with proper accounting for degradation or embrittlement that might have
taken place during component service. While all three aspects just discussed are important,
confirmation usually relies on the materials aspects gathered in a “metallurgical failure analysis.”
In this context, “failure analysis” is commonly understood to mean “metallurgical failure
analysis.” Indeed, in some instances the information gathered from the metallurgical failure
analysis is sufficient to provide a high degree of the confidence in the failure mechanism and the
other two facets of failure analysis (stress/fracture analysis, and operational influence) are not
essential.
The project was developed to extend over a two year period from 2006 - 2007and this Interim
Report details the results from the first year of research. A final report covering both years will
be completed by the end of 2007.
ix
CONTENTS
1 INTRODUCTION ....................................................................................................................1-1
1.1 References ......................................................................................................................1-3
2 SEQUENTIAL STEPS TYPICALLY INVOLVED IN A FAILURE ANALYSIS........................2-1

2.1 Collection of Failure Background Data............................................................................2-2
2.1.1 Literature Review Relevant to the Failed Component .............................................2-6
2.1.2 Literature Review Relevant to the Postulated Failure Mechanism(s) ......................2-7
2.2 Photographic Documentation and Identification..............................................................2-7
2.3 Nondestructive Examination.......................................................................................2-8
2.4 Preliminary Examination of Failed Component .............................................................2-18
2.5 Sample Removal, Identification, and Preservation of Essential Features Such as
Fracture Surfaces and Deposits..........................................................................................2-19
2.5.1 Metallographic Samples ........................................................................................2-20
2.5.2 Chemical Analysis Samples ..................................................................................2-21
2.5.3 Mechanical Testing Samples.................................................................................2-21
2.6 Metallographic Examination ..........................................................................................2-21
2.6.1 Macroscopic ..........................................................................................................2-21
2.6.1.1 Visual, Low Power Microscopic, SEM............................................................2-21
2.6.1.2 Hardness .......................................................................................................2-22
2.6.1.3 Polishing and Etching ....................................................................................2-23
2.6.1.4 Interpretation..................................................................................................2-23
2.6.2 Microscopic............................................................................................................2-24
2.6.2.1 Polishing and Etching ....................................................................................2-24
2.6.2.2 Microhardness ...............................................................................................2-26
2.6.2.3 Interpretation..................................................................................................2-27
2.7 Chemical Analysis .........................................................................................................2-28
2.7.1 Bulk........................................................................................................................2-29
2.7.2 Local ......................................................................................................................2-30
xi
2.8 Mechanical Testing .......................................................................................................2-30
2.8.1 Options Based on Material Availability and Support to Failure Type.....................2-31
2.8.2 Interpretation .........................................................................................................2-32
2.9 Identification of Failure Mechanism(s) Based on Collective Information Gathered.......2-32
2.9.1 Results are Definitive and Failure Mechanism Conclusive....................................2-36
2.9.2 Results are Inconclusive and Additional Investigation Required ...........................2-37
2.10 Synthesis of Results from Failure Analysis and Unit Operation to Establish a
Root Cause .........................................................................................................................2-37
2.11 Assemble and Organize the Evidence, Formulate Conclusions, Write Report,
Including Options to Avoid Repeat Failures ........................................................................2-39
2.12 References ..................................................................................................................2-40
3 TYPICAL LABORATORY METHODS AND EQUIPMENT UTILIZED ...................................3-1

3.1 Imaging ...........................................................................................................................3-1
3.1.1 Visual and Low Power Stereo Microscopy ..............................................................3-1
3.1.2 Optical Light Microscopy (OLM) ..............................................................................3-2
3.1.2.1 Magnification, Numerical Aperture, Resolution, and Depth of Field ..............3-14
3.1.2.2 Sample Preparation (mounting, polishing, etching) .......................................3-19
3.1.2.3 Phases and Particle Identification..................................................................3-19
3.1.2.4 Grain Size Estimates .....................................................................................3-20
3.1.3 Scanning Electron Microscopy (SEM) ...................................................................3-23
3.1.3 Scanning Electron Microscopy (SEM) ...................................................................3-23
3.1.3.1 Magnification and Resolution Relative to OLM ..............................................3-23
3.1.3.2 Sample Preparation .......................................................................................3-25
3.1.3.3 Viewing Modes ..............................................................................................3-26
3.1.3.3.1 Polished or Polished and Etched ...........................................................3-26
3.1.3.3.2 Fracture Surfaces...................................................................................3-26
3.1.4 Transmission Electron Microscopy........................................................................3-26
3.1.4.1 Magnification and Resolution Relative to OLM and SEM ..............................3-27
3.1.4.2 Types of Samples and Sample Preparation ..................................................3-30
3.1.4.3 Viewing Modes ..............................................................................................3-33
3.2 Chemical Analysis .........................................................................................................3-34
3.2.1 Bulk (Macro) ..........................................................................................................3-34
3.2.1.1 Conventional Quantitative Analysis – “Wet Chemical Analysis” ....................3-34
3.2.1.2 Spectrographic...............................................................................................3-34
xii
3.2.1.2.1 Optical Emission ....................................................................................3-34
3.2.1.2.2 X-Ray Techniques (Energy and Wave Length Approaches) .................3-34
3.2.1.3 Combustion and Other Methods (Carbon, Sulfur, Nitrogen, Hydrogen,
Oxygen) .....................................................................................................................3-34
3.2.2 Local and Residual Level (Micro) ..........................................................................3-34
3.2.2.1 Spectrographic Using SEM or Microprobe (Energy Dispersive and
Wavelength Including Detectability and Accuracy) ....................................................3-34
3.2.2.2 Sample Dissolution and “Plasma” Detection Techniques (Flame,
Inductively Coupled Plasma) .....................................................................................3-34
3.2.3 Particle (Compound) Identification ........................................................................3-34
3.2.3.1 X-Ray Diffraction............................................................................................3-34
3.2.3.2 Electron Diffraction (TEM)..............................................................................3-34
3.3 Mechanical Testing .......................................................................................................3-34
3.3.1 Hardness ...............................................................................................................3-34
3.3.1.1 Macro .............................................................................................................3-34
3.3.1.2 Micro ..............................................................................................................3-35
3.3.2 Tensile ...................................................................................................................3-35
3.3.3 Fracture .................................................................................................................3-35
3.3.3.1 Charpy V-Notch .............................................................................................3-35
3.3.3.2 Plane Strain Fracture Toughness (KIc)..........................................................3-35
3.3.3.3 Elastic-Plastic Fracture Toughness (JIc) .......................................................3-35
3.3.3.4 Other Measure of Fracture Resistance..........................................................3-35
3.3.4 Creep-Rupture.......................................................................................................3-35
3.3.5 Fatigue...................................................................................................................3-35
3.3.6 Small Sample Testing Options ..............................................................................3-35
3.4 References ....................................................................................................................3-35
4 SPECIAL TECHNIQUES........................................................................................................4-1
To Be Added in 2007.............................................................................................................4-1
5 TYPICAL MATERIALS OF CONSTRUCTION.......................................................................5-1

To Be Added in 2007.............................................................................................................5-1
6 DAMAGE MECHANISMS ......................................................................................................6-1

To Be Added in 2007.............................................................................................................6-1
7 CHARACTERISTIC DAMAGE FOR SPECIFIC COMPONENTS ..........................................7-1
xiii
To Be Added in 2007.............................................................................................................7-1
8 SUMMARY, CONCLUSIONS, RECOMMENDATIONS .........................................................8-1

To Be Added in 2007.............................................................................................................8-1
A COMMONLY USED ETCHANTS AND PROCEDURES ...................................................... A-1

A.1 Macroetches .................................................................................................................. A-1
A.1.1 Ferritic Steels ......................................................................................................... A-1
A.1.2 Stainless Steels...................................................................................................... A-4
A.2 Microetchs ..................................................................................................................... A-5
A.2.1 Ferritic Steels ......................................................................................................... A-5
A.2.2 Stainless Steels...................................................................................................... A-6
A.3 References..................................................................................................................... A-7
xiv
LIST OF FIGURES
Figure 2-1 Illustration of the Various Disciplines and Organizations Which May be
Involved when a Component Fails [2-1].............................................................................2-2
Figure 2-2 DeLong (FN) Diagram for Estimating the Ferrite in Stainless Steel Weld Metal
[2-15] ................................................................................................................................2-17
Figure 2-3 WRC 1988 (FN) Diagram for Estimating the Ferrite in Stainless Steel Weld
Metal [2-15] ......................................................................................................................2-18
Figure 2-4 Indentations made by Knoop and Vickers Indenters in the Same Work Metal
at the Same Load [2-17]...................................................................................................2-27
Figure 3-1. Illustration of a Stereo Zoom Microscope Mounted on a Boom Arm as
Typically Used for Low Power Three-Dimensional Viewing of Failed Components...........3-2
Figure 3-2 Diagrammatic illustration of a Transmitted Light Compound Microscope in
Which the Intermediate Image formed by the Objective and Tube Lens is Enlarged
by the Eyepiece [3-2] .........................................................................................................3-3
Figure 3-3 Photograph of a Typical Bench-Type Reflected Light Compound Microscope
Used in Metallography [3-3] ...............................................................................................3-4
Figure 3-4 Illustration of the Reflected Light Microscope and the Trace of Rays through
the Optical Lens System [3-4] ............................................................................................3-5
Figure 3-5 Photomicrographs of Ferrite at 200X. The Specimen was Mechanically
Polished and Deep Etched [3-5] ........................................................................................3-8
Figure 3-6 Photomicrographs of Cast Alloy at 200X.The Specimen was Mechanically
Polished but was not Etched [3-5]......................................................................................3-9
Figure 3-7 Photomicrographs of Cadmium at 100X. The Specimen was Chemically
Polished [3-5] ...................................................................................................................3-10
Figure 3-8 Photomicrographs of High Manganese Cast Steel at 125X. The Specimen
was Mechanically Polished but was not Etched [3-5].......................................................3-11
Figure 3-9 Photomicrographs of Beta Brass at 400X. The Specimen was Mechanically
Polished and Etched [3-5] ................................................................................................3-12
Figure 3-10 Photomicrographs of Zircaloy at 400X. The Specimen was Chemically
Polished [3-5] ...................................................................................................................3-13
Figure 3-11 Photomicrographs of Zirconium at 400X. The Specimen was Chemically
Polished [3-5] ...................................................................................................................3-14
Figure 3-12 Portion of the Electromagnetic Spectrum Occupied by Visible Light [3-7]............3-16
Figure 3-13 A Diagrammatic Representation of a Scanning Electron Microscope [3-13] ........3-24
Figure 3-14 Illustration of the Emissions Produced by from the Sample in Response to
the Incident Electron Beam in the SEM [3-13] .................................................................3-25
xv
Figure 3-15 Comparison of the Imaging System for a Transmission Electron Microscope
and a Transmitted Light Optical Microscope [3-14]..........................................................3-28
Figure 3-16 Steps in Preparing an Indirect Carbon Replica using a Cellulose Acetate or
Polyvinyl Alcohol (PVA) Primary Replica [3-13] ...............................................................3-32
Figure 3-17 Schematic illustration of the Preparation of an Extraction Replica Used to
Extract Particles for Analysis in the TEM [3-13] ...............................................................3-33
xvi
LIST OF TABLES
Table 2-1 ASTM and ASME Standards for Nondestructive Examination Methods [2-4] ...........2-9
Table 2-2 Nondestructive Examination Methods to Detect Defects in Welds and
Components [2-5].............................................................................................................2-12
Table 2-3 A Partial Compilation of Failure Mechanisms for Boiler Tubes [2-26 – 2-28] ..........2-33
Table 3-1 Best Achievable Resolution and Associated Magnification for Optical Light
Microscopy .......................................................................................................................3-17
Table 3-2 Relationship between ASTM Grain Size Number and the Average Grain
Diameter...........................................................................................................................3-21
Table 3-3 Simplified Summary of Grain-Size Effects on Properties of Steels..........................3-22
Table 3-4 Comparison of Characteristics of Optical Microscopes, Transmission Electron
Microscopes, and Scanning Electron Microscopes [3-13]................................................3-29
Table 3-5 Comparison of Advantages and Disadvantages of Optical Microscopes,
Transmission Electron Microscopes, and Scanning Electron Microscopes [3-13] ...........3-30
xvii
1
INTRODUCTION
Failures of fossil power plant components have to be viewed in terms of the design basis,
historical expectations, and consequences of failure. A multiplicity of design bases exists
internationally, but they all rest on the foundation of avoiding gross distortion and providing a
reasonable margin against bursting.
In terms of ASME Section I, a simplified “design by rule” approach is used with the Foreword
stating the following: “the objective of the rules is to afford reasonably certain protection of life
and property and to provide a margin for deterioration in service so as to give a reasonably long,
safe period of usefulness.” It should be noted that explicit times or number of cycles are not
defined in the words “reasonably long, safe period of usefulness” and the historical expectation
then becomes a basis for deciding whether a particular failure is “expected” or “not expected.”
Indeed, some components have an expectation of essentially infinite life whereas other
components have an implied finite life.
Similar to ASME, fossil plant power plant design in other countries is predominantly done on a
design-by-rule basis rather than a design-by-analysis basis. However, accepted analytical
methodologies such as finite element stress analysis and fracture mechanics analysis is more
broadly integrated into the design and fabrication process for International Codes than for ASME
Section I. In contrast to ASME Section I, some International Codes have a formal requirement
for fatigue analyses. They additionally have formal requirements to demonstrate that hot tensile
properties are “guaranteed” whereas ASME Section I implements a trend-curve approach based
on conformance to specified room temperature strength which does not guarantee specific
compliance for hot tensile strength but is a rational engineering approach in which the
uncertainties in strength are covered by the design margins.
In most European Codes there are formal requirements for impact strength (fracture toughness)
in base materials and weld metals. In ASME similar requirements for base metals exist in some
of the material specifications but there are no counterpart requirements for weld metals. Finally,
many International Codes have mandated in-service inspection (ISI) requirements, and extension
of the interval of operation is based on the outcome of the examinations performed. Some
International Codes even establish a specific “retirement date” for power plant components.
While ASME presently has no such counterpart rules, a Post-Construction Committee has been
operative for several years and Post-Construction rules are anticipated.
Failure can usually be classified as: (a) loss of function without associated safety or severe
secondary damage implications, or (b) loss of function in which there are safety implications or
an associated high probability of secondary damage. An example of the former is a leak in a
reheater tube which is not severe enough to immediately bring the unit offline and which would
be contained within the boiler setting, even if a catastrophic rupture were to occur. An example
1-1
Introduction
of the latter is creep-rupture in a long axial crack in seam welded superheater link piping where a
massive steam leak and long crack extension occurs suddenly.
Failures in components of fossil plant boilers, piping, and turbines are common occurrences and
each encounter raises the issue of whether a “failure analysis” should be performed and, if so,
what are the essential elements of that failure analysis. A comprehensive failure analysis has
three major facets that, when combined, provides a consistent framework to claim a “total
understanding.” These three aspects are: (a) stress and fracture mechanics analysis based on
both design and actual operation, (b) confirmation of actual operation relative to the design
assumptions and identification of departures that are relevant to the failure, and (c) the
mechanical and environmental behavior of materials under both the design scenario and actual-
operations scenario with proper accounting for degradation or embrittlement that might have
taken place during component service. While all three aspects just discussed are important,
confirmation usually relies on the materials aspects gathered in a “metallurgical failure analysis.”
In this context, “failure analysis” is commonly understood to mean “metallurgical failure
analysis.” Indeed, in some instances the information gathered from the metallurgical failure
analysis is sufficient to provide a high degree of the confidence in the failure mechanism and the
other two facets of failure analysis (stress/fracture analysis, and operational influence) are not
essential.
Thus, the emphasis of this research is the development of a “fossil power plant components
failure analysis guideline” which emphasizes metallurgical failure analysis.
The first four Chapters identify the instruments, equipment, and procedures which are utilized to
perform failure analyses as well as the sequential steps to be followed. The remaining Chapters
identify typical materials of construction, generic damage mechanisms, and characteristic
damage for an assortment of power plant components.
In closing this introduction, some quotes from the marvelous book by Henry Petroski, To
Engineer is Human: The Role of Failure in Successful Design, will be given [1-1]:
“I believe that the concept of failure -- mechanical and structural failure in the context of
this discussion -- is central to understanding engineering, for engineering design has as
it’s first and foremost objective the obviation of failure. Thus the colossal disasters that
do occur are ultimately failures of design, but the lessons learned from those disasters can
do more to advance engineering knowledge than all the successful machines and
structures in the world. Indeed, failures appear to be inevitable in the wake of prolonged
success, which encourages lower margins of safety. Failures in turn lead to greater safety
margins and, hence, new periods of success. To understand what engineering is and what
1-2
Introduction
engineers do is to understand how failures can happen and how they can contribute more
than successes to advance technology.”
“All the successes of engineering, as far back in history as the pyramids and as far into
the future as the wildest conceptions of mile-high skyscrapers, may be imagined to have
begun with a wish to achieve something without failure, where "without failure" to the
engineer means not only to stand without falling down but also to endure with what
might be called "structural soundness." Unsound structures -- those that are eaten away
by rapid corrosion, those that have repeated service breakdowns under ordinary
conditions, those that suffer from fatigue cracking after not so many years of use -- may
be thought to have been failures as surely as if they had collapsed during construction.
And no matter how ingenious or attractive his conception may appear in his imagination
or on paper, if a designer overlooks just one way in which his structure may fail, all may
be for naught.”
Thus, a central tenet of Petroski is that engineers learn more from failures that from successes
and that the insights gleaned from failures feeds back into the knowledge base that is essential to
avoid repetitions of similar failures and to improve future designs. It is with this broader
perspective that failure analysis should be viewed.
1.1 References
1-1. Henry Petroski, To Engineer is Human: The Role of Failure in Successful Design, Barnes
and Noble Books, New York, 1994.
1-3
2
SEQUENTIAL STEPS TYPICALLY INVOLVED IN A
FAILURE ANALYSIS
The eleven steps typically involved in a failure analysis are as follows:

• Collection of Failure Background Data
• Photographic Documentation and Identification
• Nondestructive Examination (Visual, MT, PT, UT, ET, Ferrite Measurement, Other)
• Preliminary Examination of Failed Component
• Sample Removal, Identification, and Preservation of Essential Features Such as Fracture
Surfaces and Deposits
• Metallographic Examination
• Chemical Analysis
• Mechanical Testing
• Identification of Failure Mechanism(s) Based on Collective Information Gathered
• Synthesis of Results from Failure Analysis and Unit Operation to Establish a Root Cause
• Assembly and Organize the Evidence, Formulate Conclusions, Write Report, Including
Options to Avoid Repeat Failures
Conceptually, these steps can be thought to occur sequentially, but some may proceed in parallel,
or in an order differ from that shown. Whether or not the steps occur sequentially, ideally all
steps should be performed. However, in particular circumstances some of the steps may be
eliminated if the evidence already compiled makes a compelling case. If steps are eliminated,
the investigator should be aware of the pitfalls that exist with “incomplete data” and be prepared
to rationalize the abbreviated investigation.
A disciplined approach, detached from emotional involvement, should be taken in all failure
analyses to maximize the possibility of “arriving at the truth.” There is probably nothing more
crucial than time spent on the front end to develop a plan and to gain “buy in” by all involved
parties. While detachment from emotional involvement is an ideal, it must be appreciated that
one of the outcomes from a failure analysis may be the assignment of blame, either to
individuals, organizations, or a system of management. Such assignment must be done with
diplomacy and discretion since an accusatory environment during the investigation can close
lines of communication and impede the process.
2-1
Sequential Steps Typically Involved in a Failure Analysis
2.1 Collection of Failure Background Data
The complexity of power house equipment dictates that a myriad of disciplines are involved
from conception to commissioning, and then in operation, maintenance, repair/replacement and
perhaps eventual retirement of the equipment. All components have a unique strategy in this
“birth-to-death” design/build/operate cycle. The utility seeks to strike a proper balance between
cost and reliability while assuring personnel safety.
A conceptual framework showing the various disciplines and organizations which may be
involved in a component failure analysis is shown in Figure 2-1 (adapted from Reference 2-1).
Design Operations Engineering
Component
Failure
Fabrication Maintenance Management
Figure 2-1
Illustration of the Various Disciplines and Organizations Which May be Involved when a
Component Fails [2-1].
Even though all disciplines or organizations illustrated in Figure 2-1 may not get involved in
each failure analysis, it is important at the outset to recognize that multiple disciplines may be
needed to gain a full understanding. Exclusion or inclusion of disciplines or organizations
relative to the failure investigation should be done as deliberate actions rather than as defaults.
A brief summary of some of the background data which can relate to any of the six disciplines follows.
DESIGN
• Is the component designed to an industry code or standard, or to the proprietary practices of
an OEM?
• Can the design basis and the original calculations for component sizing be obtained from
historical records or inferred from “back calculations?” Critical items are design temperature
and loads (pressure).
• Was the material of construction procured from a recognized industry standard such as
ASTM, ASME, Euronorm, British Standard, AFNOR, or DIN, or is it a proprietary material?
Have there been material changes during the component life cycle?
• What are the critical component dimensions, tolerances, and surface finishes for the original
design and are there drawings documenting these? Have there been changes during the
component life cycle?
• Did the design formally account for environmental effects (e.g., a corrosion allowance)? If
so, does experience demonstrate proper accounting for environment effects?
2-2
• Did the design formally account for weld or casting quality and is there evidence of
compliance to an associated Code requirement?
• What are the global environmental considerations (fuels, combustion products, fluids, gases,
flow rates, surface protection, etc.) and have these changed during the period of service?
• Does the component have overpressure or over temperature protection or alarms? What are
the bases for their settings and the consequences of exceeding limits?
FABRICATION & ERECTION

• Was cold-forming involved in the fabrication or erection of the component? Were limits
placed on cold-forming deformation (strains) and is there direct or inferential evidence that
these were adhered to?
• Was hot-forming involved in the fabrication or erection of the component? Were limits
placed on hot forming deformation (strains) and is there direct or inferential evidence that
these were adhered to?
• What heat treatments were involved in the fabrication and were these mandated by Code
rules? Is there documentation of compliance with the heat treatment rules during
fabrication? Have the heat treatment rules been modified since initial fabrication and, if so,
does this have relevance?
• Was nondestructive examination mandated for the component (welds in particular) during
initial construction? What type? Were findings reported in detail or more on the basis of
pass-fail?
• Were field modifications made during initial installation or at any later date? Are these
documented?
• Did the fabrication and erection include measures to avoid excessive residual stresses (such
as stress relief)? Is there any evidence that excessive residual stresses played a role in
component failure?
• Were formal measures implemented during fabrication and erection to assure component
cleanliness and freedom from blockage of internal passages?
MAINTENANCE
• What were the maintenance practices applicable to the failed component in the time period
preceding failure?
• Did the component have “fouling limits” that were monitored continuously or periodically,
the violation of which resulted in cleaning actions (e.g., deposits on the ID surface of
waterwall tubes)?
• Were major cleaning actions, such as acid cleaning of the boiler, performed during outages?
Were these performed as-intended or was there evidence that overly aggressive cleaning may
have occurred?
2-3
• Were repairs performed on the failed component during any time prior to failure? What was
the nature and frequency of repairs? Were the repairs ideally executed or might some
compromises have been made?
• Were condition assessments performed for the component during any periods preceding
failure? Results?
OPERATIONS
• During a cold start, what are the intended temperature and pressure ramp rates and the
associated tolerances on each?
• During a warm (hot) start, what are the intended temperature and pressure ramp rates and the
associated tolerances on each?
• During a normal shutdown, what are the intended temperature and pressure ramp rates and
the associated tolerances on each?
• During an emergency shutdown, what are the intended temperature and pressure ramp rates
and the associated tolerances on each?
• During a unit trip, what is the extent to which temperature and pressure ramp rates can be
controlled?
• Which boiler components limit the startup ramp rate(s) and what is the associated limit?
• Which boiler components limit the shutdown ramp rate(s) and what is the associated limit?
• What limits the startup ramp rate(s) for the turbine and what is the associated limit?
• What limits the shutdown ramp rate(s) for the turbine and what is the associated limit?
• In the total system, what limits the startup and shutdown time and what are the associated
ramp rates?
• Are there “hold times” to “clean up” the water chemistry? When do they occur? What
particular contaminants (dissolved oxygen, silica, carryover of solids, etc.) restrict startup
time?
• Is there a data acquisition system which monitors megawatts, turbine speed(s),
turbine/generator vibration, critical temperatures, pressures, flow rates, and water chemistry?
For what time period can such data be retrieved?
• What are the protective measures taken during shutdown and layup to minimize out-of-
service corrosion damage?
• What restricts the full load capability of the unit?
• At full load, does the unit achieve “nameplate rating?”
• Has there been a decline in full load capability and are the reasons for the decline known?
• Since the unit was commissioned, what is the number of: cold starts, warm starts, unit trips,
and emergency shutdowns?
• What are the cumulative operating hours?
2-4
• Are there histograms of: annual operating hours, annual generation lost due to forced
outages, planned outages, and unit available but not needed?
• In a typical year, are there histograms of: megawatts, main steam outlet temperature and
pressure, hot reheat steam temperature and pressure?
• At full load steady state conditions, do the temperatures and pressure remain relatively
constant or do they cycle?
• Is the unit considered to be base loaded or cycling?
• For cycling, what particular cycles are experienced in terms of turndown, two-shifting, week-
end offline, etc.?
• What fuels were used as the design basis for the unit?
• Over the period of operation, have there been fuel switches and what is the current fuel?
• Have pulverizers, burners, and air introduction to the boiler been modified from the original
design and how has this impacted unit operation?
• Has “back end” air quality control equipment (precipitators, scrubbers, selective catalytic
converters) been added and how has this impacted unit operation?
• As a generalization, has the unit been operated within the “design envelop?” If not, what
sorts of violations have occurred and what equipment might be adversely impacted?
Without doubt, unit operation plays a major role in determining availability and thus in
minimizing failures. Failures are often attributable to “abnormal operation” but the specific
connection is often elusive. Furthermore, metallurgist and engineers performing failure analysis
often are not well informed about operations and have difficulty accessing information to
establish “operational cause and effect.”
The preceding list of questions regarding operations should be reviewed and those questions that
have relevance to a particular failure should be pursued. In some cases, a question may not
relate to the current failure, but it is good general information to gather and update periodically
for each unit in the system.
ENGINEERING
The person with principal responsibility to lead a failure investigation and to develop mitigating
strategies typically resides in the engineering organization and is commonly a metallurgist. But
other engineering disciplines may take the lead role in particular situations, especially if they
additionally have metallurgical expertise or have a metallurgist on the investigation team. Some
of the traditional roles of engineering follow:
• What are the damage (degradation) mechanisms for the failed component?
• Of the several embrittlement mechanisms [2-3], is one likely to be involved in the current
failure?
2-5
• Is there an inherent design or materials deficiency associated with the failed component and
are there pragmatic corrective actions, redesigns, or material substitutions that can be taken
in the available period before returning the unit to service? At a later time?
• Is the failure confined to only one component or might there be other similar components or
locations that have suffered similar damage?
• Has the damage extent been adequately established and will the corrective actions mitigate
future failures?
• Are there associated plant maintenance and operation guidelines that require updating in
response to the failure investigation?
MANAGEMENT
• Did a deficient purchase specification play a role in the failure? If so, what revisions should
be made? How will these revisions impact cost, availably, and delivery for future
procurement?
• Does the failure dictate changes to corporate policies for operations, maintenance, and
engineering? If so, how will these be implemented?
• Did the role of unit dispatch in the operation of the utility contribute to the failure? If so,
what remedial options exist?
• Is the failure uniquely confined to a specific OEM and have they been appropriately involved
in diagnosis and resolution? Has the OEM issued “industry alerts” warning users of the
issues?
• As a consequence of the failure, should the unit operational mode (base load, load following,
two-shifting, etc.) or the order of dispatch be changed? Would such change adversely impact
other units?
• As a consequence of the failure, what outage planning should be implemented for the
specific unit and other units in the system?
2.1.1 Literature Review Relevant to the Failed Component
There are many ways to organize the major systems in a power plant. In the broadest sense, they
may be organized as mechanical and electrical.
The major mechanical systems are typically the boiler and auxiliaries and the turbine and
auxiliaries. Boiler auxiliaries typically include the components from the turbine exhaust to the
economizer inlet and the fuel and air handling systems. Turbine auxiliaries include the valves,
oil pumps, filters, and associated piping.
The major electrical systems include the generator and associated auxiliaries, exciter,
transformers, and control systems.
Information that is useful to a failure investigation is often organized as “component specific”

and may be accessed through Internet Information Systems such as EPRI.com. On occasion a
2-6
“Google Search” may yield useful information. Technical journals that focus on the power
sector often have relevant information. Textbooks, regardless of vintage, are frequently valuable
resources. Finally, communications with engineers within your utility or those in a similar
position in other utilities is usually beneficial.
2.1.2 Literature Review Relevant to the Postulated Failure Mechanism(s)
Failure mechanisms are typically discussed in metallurgical and mechanical engineering

literature. Mechanical engineering and mechanical metallurgy typically cover damage due to
loads and load sequences (i.e., fatigue) while physical metallurgy seeks to establish the
relationship between alloying elements, phases/microstructure, heat treatment, and the material
behavior, including environmental effects.
Whether organized as component specific or by failure mechanism, similar options to those

discussed in the previous section are available.
2.2 Photographic Documentation and Identification
In former years, photography was film-based and there was no instant validation that the desired
pictures had, indeed, been captured. This limitation placed a premium on use of professional
photographers having a collection of cameras, lens, and lighting equipment to accommodate any
situation. But the advent of digital photography has dramatically changed the situation and is
now the mainstay in photographic documentation.
The discussion in this section is mainly confined to photography that occurs at the failure
location. Considering the modest cost, photographic documentation for a failure analysis should
not be compromised and it is better to err on the side of excess. To the extent possible,
photographs should be taken of the failure and a broad area near the failure site prior to anyone
disturbing the surroundings. There is often collateral damage and secondary damage, not
attributable to the initial failure, which may be hard to distinguish from the initial failure when
the area is first viewed. To maintain relative size, a ruler or object(s) of known size are best
included in the field of view.
If there is wreckage involving a scattering of pieces, care should be taken to document each
piece in the as-found location prior to moving it.
Ideally the camera will have a collection of lens or a versatile telephoto lens, flash, and sufficient
resolution (4 mega pixels or greater) to capture the desired images and allow for enlargement of
selected regions by 2X to 5X without losing appreciable detail. Color photography is advisable
and the “native resolution” of the camera should be maintained in the initial photography.
Compression to reduce the size of the photographic images should be deferred until the final
stages since it potentially compromises image detail.
The photography should proceed from the overview (far away) to increasing closeness to the
failure itself. It should capture the relative location of non-failed portions of the component. If
2-7
the design has similar replicate locations, photographs should be taken of these for comparison to
the failed location.
Prior to moving the failed component and identification scheme should be adopted to facilitate
“mental reassembly” of the pieces as the samples are removed for further investigation.
Where fractures are visible in the field examination, general photography should be performed of
the entire fractured part, including broken pieces, to record their size and condition and to show
how the fracture is related to the components of the part. The fracture surfaces should be
illuminated by direct lighting and by oblique lighting to fully image the relevant features.
2.3 Nondestructive Examination
Nondestructive examination (NDE) is very important in the life-cycle management of power

plant components. During the procurement phase of new or replacement components, NDE is
applied to assure that the component quality conforms to the procurement specification and the
construction code. During component service, NDE is often applied to perform condition
assessments and remaining life estimation. Following a failure, NDE is used to ascertain
whether the failure extends beyond that which is visually observed and whether similar
components/locations in the same unit or replicate units display early signs of distress.
The type of NDE that is performed in the field prior to removal of the failed component requires
careful consideration in terms of altering the evidence or introducing chemicals that may
compromise certain future examinations. For example, if deposits or scale are to be studied later,
retaining them will likely impede application of some NDE methods. Additionally, the
introduction of solvents and dyes for wet fluorescent magnetic particle examination or for
penetrant testing will “wick” into cracks and “contaminate” the sample.
All International boiler and pressure vessel Codes have associated references to NDE methods
and acceptance criteria for defects. For ASME Code Construction, the NDE methods are
contained in Section V (essentially identical to ASTM standards) and the acceptance criteria are
either in the Construction Code (i.e., Sections I, VIII, Div. 1, or B31.1) or the product
procurement specifications in Section II, Parts A and B. The acceptance/rejection basis usually
consists of subjective (i.e., shall be free from injurious defects and shall have a workmanship
type finish) and quantitative (i.e., shall be free of relevant linear indications – length > 1/16 in
and length > three time width) criteria and are typically product specific.
The NDE methods are intended to detect surface and internal (volumetric) discontinuities in base
materials, welds, and components. The skill, experience, and integrity of the inspection
personnel are essential to obtain meaningful results.
There are several ASTM, ASME, and DIN EN standard practices and methods for the various
NDE techniques. These standards provide definitions of terms, recommended practices,
reference photographs, and quality control requirements. Table 2-1 lists the applicable standard
by number and title for some of the NDE methods to be discussed in this section [2-4].
2-8
Table 2-1
ASTM and ASME Standards for Nondestructive Examination Methods [2-4]
Liquid Penetrant Testing

ASTM E 165 Standard Practice for Liquid Penetrant Inspection Method
ASTM E 270 Standard Definitions of Terms Related to Liquid Penetrant Inspection
ASTM E 433 Standard Reference Photographs for Liquid Penetrant Inspection
ASME Section V Article 6
Nondestructive Examination of Welds; Penetrant Testing of Welds;
DIN EN 1289 Acceptance Levels
Magnetic Particle Testing

ASTM E 709 Standard Recommended Practice for Magnetic Particle Examination
ASTM E 269 Standard Definitions of Terms Relating to Magnetic Particle Examination
Nondestructive Examination of Welds; Magnetic Particle Testing of Welds;
DIN EN 1291 Acceptance Levels
Radiographic Testing
ASTM E 94 Standard Practice for Radiographic Testing
ASTM E 142 Standard Method for Controlling Quality of Radiographic Testing
ASTM E 586 Standard Definitions of Terms Relating to Gamma and X Radiography
Nondestructive Examination of Welds; Radiographic Examination of
DIN EN 12517 Welded Joints; Acceptance Levels
Ultrasonic Testing
ASTM E 500 Standard Definitions of Terms Relating to Ultrasonic Testing
Standard Recommendation Practice for Fabrication and Control of Steel
ASTM E 428 Reference Blocks Used in Ultrasonic Inspection
ASTM E 213 Standard Practice for Ultrasonic Inspection of Metal Pipe and Tubing
Standard Practice for Ultrasonic Contact Examination of Weldments Pulse-
ASTM E 164 Echo Contact Method
Standard Recommendation Practice for Ultrasonic Pulse-Echo Straight-
ASTM E 114 Beam Testing by Contact Method
Nondestructive Testing -- Ultrasonic Examination - Part 1: General
DIN EN 583-1 Principles
Nondestructive Examination of Welds; Ultrasonic Examination;
DIN EN 1713 Characterization of Indications in Welds
Testing Agency Qualification

Standard Practice for Determining the Qualification of Nondestructive
ASTM E 543 Testing Agencies
Nondestructive Testing -- Qualification and Certification of NDT Personnel -
DIN EN 473 General Principles
2-9
Commonly used nondestructive methods of examination of failed components include:

• Visual examination (VT)
• Ultrasonic examination (UT)
• Magnetic particle testing (MT)
• Liquid penetrant testing (PT)
• Radiography (RT)
• Eddy current testing (ET)
• Portable alloy analysis
• Plastic replication
• Ferrite measurement
The examination methods are complementary and a combination of methods will often be used
to gain more complete information. The equipment needs, application, advantage, and
limitations of the most common methods are shown in Table 2-2 [2-5].
Visual inspection, often aided by MT or PT, is used to detect surface cracks, which appear as
thin wavy or broken lines. Subsurface defects can be detected using either UT or RT. Pulse-
echo UT instruments send out pulses of high-frequency sound, which penetrates the area to be
inspected. Pulses reflecting back to the UT transducer indicate locations and dimensions of
defects. Ultrasonic equipment can also be used to measure original and remaining wall
thickness. Isotope-powered gamma ray cameras are used in radiography to provide a permanent
record of inspection.
The NDE technique(s) selected should be the most effective in locating the types of
discontinuities likely to be encountered. Using the same technique for every component or weld
inspection can result in missing significant defect indications. Weld accessibility, material and
weld type, and NDE method detection capabilities and limitations should determine which
techniques are best suited for the application. Examples provided in the following sections, that
are typical of boiler and piping components, show the capability of the different methods [2-6].
Visual Inspection
Visual inspection is at the heart of a failure investigation because it provides the initial basis to
plan subsequent work. This method is used for the detection of cracks that are open to the
surface (either partial or through-wall), crack paths and proximity for multiple cracks, and for
preliminary observations of the fracture surface. In boiler tubes, VT is used to judge fireside
flyash erosion and corrosion problems and broken hangers and attachments. Borescopes, fiber
optics, mirrors, and miniature closed circuit television systems have been used to inspect the ID
of headers (for example, for ligament cracking in economizer inlet headers and secondary
superheater outlet headers). Visual examination is also important in welding to check for correct
weld fit-up, cracks, underfill, incomplete penetration, cold laps, surface porosity and gas pockets,
arc strikes, weld spatter, and undercut.
2-10
Ultrasonic Testing
UT can provide significant data about tube wall thickness and subsurface material condition
including cracking. Thickness measurements will provide information as to whether any tubing
or thick-walled components are below the minimum Code requirements and, thus, need
evaluation, weld repair, or replacement. Periodic thickness surveys in the same locations will
indicate the thinning rate and facilitate estimates for the timing of pressure part replacements.
In the broad sense, ultrasonic testing is usually done with longitudinal waves (straight beam) or
shear waves (angle beam). Longitudinal waves are mainly used to check for thickness, for
subsurface laps and laminations that are parallel to the surface, or for other similar smaller
2-11
Table 2-2
Nondestructive Examination Methods to Detect Defects in Welds and Components [2-5]
defects such as clad disbond. Shear waves are used to detect cracks and establish their size and
orientation. Historically shear wave inspections have been performed using fixed angle
transducers with multiple transducers often required to fully explore the region of interest. A
recent innovation called linear phased arrays uses a collection of transducers in a single head
that are be pulsed in a manner to sweep through various angles without moving the head and to
enable a degree of focusing.
2-12
Ultrasonics is also used to check for certain types of in-service damage and to ensure the
integrity of repair welds. Ultrasonic techniques have been developed for the detection of
hydrogen damage [2-7] and corrosion fatigue [2-8].
Magnetic Particle Testing
The magnetic particle method is based on the principal that magnetic field lines, when present in
a ferromagnetic material, will be distorted by a change in material continuity, such as a sharp
dimensional change or a discontinuity (crack). If the discontinuity is open to or close to the
surface of a magnetized material, flux lines will be distorted at the surface, a condition termed as
“flux leakage.” When fine magnetic particles are distributed over the area of the discontinuity
while the flux leakage exists, they will be held in place and the accumulation of particles will be
visible under proper lighting conditions. Thus, MT is used for locating surface or near-surface
discontinuities in ferromagnetic materials. Typical discontinuities detectable by this method are
cracks, seams, laps, cold shuts, and laminations
The four basic type of current used in magnetic particle examination to establish magnetization
of the part are alternating current (AC), single phase half-wave rectified alternating current, full-
wave rectified alternating current, and direct current (DC). Associated with AC is a “skin effect”
that confines the magnetic field at or near the surface whereas rectified AC and DC produce a
more penetrating field in the part that is capable of detecting near-surface discontinuities. The
magnetic field is introduced though yokes, prods, or wire wraps.
Both wet and dry particles are used in MT. The most frequently employed colors for dry
powders are light grey, black, red, or yellow. On occasion the part is painted with a color
(usually white) that contrasts with the powder color to facilitate viewing and photography.
Special dry powders are available for use at temperatures up to 600°F (315°C) that are especially
useful for in-process monitoring of weld repairs performed with preheat. Most often wet
fluorescent magnetic particles are designed to be suspended in a vehicle such as water or light
petroleum distillate and are viewed under black (ultraviolet) light, causing the particles to
fluoresce.
MT is widely used to detect defects inside headers and drums, at any stress raisers, such as at the
toes of fillet welds, or in the blade attachment region of steam turbine rims. Magnetic particle
inspection is performed to ensure that all defects are removed prior to welding, during welding,
and on the completed weld.
Liquid Penetrant Testing
The liquid penetrant examination method (PT) is an effective means for detecting discontinuities
which are open to the surface of nonporous metals and other materials. Typical discontinuities
detectable by this method are cracks, seams, laps, cold shuts, laminations, and porosity.
In principle, a liquid penetrant is applied to the surface to be examined and allowed sufficient
soak time to enter discontinuities. All excess penetrant is then removed, the part is dried, and a
developer is applied. The developer functions both as a blotter to absorb penetrant that has been
2-13
trapped in discontinuities, and as a contrasting background to enhance the visibility of penetrant

indications. The dyes in penetrants are either color contrast (visible under white light) or
fluorescent (visible under ultraviolet light).
PT is used for locating defects that are open to the surface. It is used in similar applications as
MT but it should not be used to check welds that are still hot. PT is also used to inspect
nonmagnetic materials (such as copper alloys or austenitic stainless steels) that cannot be
examined by the MT method. PT is also applicable when nonmagnetic and magnetic
components are joined by dissimilar metal welds. PT is a time-consuming process (nominally
15-20 minute cycle time) and is usually only applied for nonmagnetic materials.
Radiography
Two types of electromagnetic radiation are used in radiographic inspections: x-rays and gamma-
rays ( -rays). X-rays and -rays differ from other types of electromagnetic radiation (such as
visible light, microwaves, and radio waves) only in their wavelength, although there is not
always a distinct transition from one type of electromagnetic radiation to another. Only x-rays
and -rays, because of their relatively short wavelength (high energies), have the capability of
penetrating opaque materials to reveal internal flaws.
X-rays and -rays are physically indistinguishable; they differ only in the manner in which they
are produced. X-rays are produced in an x-ray tube and result from the interaction between a
rapidly moving stream of electrons and the atoms in a solid target bombarded by the electrons,
while -rays are emitted during the radioactive decay of unstable atomic nuclei (radioactive
isotopes).
Radiography (RT) is used to detect subsurface defects that are aligned in the same direction as
the radiation beam. This means that the ability of the process to detect planar discontinuities
such as cracks depends on proper orientation of the test piece during inspection. Discontinuities
such as voids and inclusions, which have measurable thickness in all directions, can be detected
as long as they are not too small in relation to section thickness. In general, features that exhibit
a 1% or more difference in absorption compared to the surrounding material can be detected
[2-9].
RT examinations can be performed for the detection of dissimilar metal weld cracking, hydrogen
damage, and corrosion fatigue. It is also used to detect service-induced defects in welds, for
examination of repair welds, and to test the welder's performance qualification test coupons.
Eddy Current Testing
Eddy current testing (ET) can be carried out on any conductive material. A probe supplied with
a high-frequency current induces an electric field in the part; the field changes in the presence of
surface or near-surface defects. These changes show up on instruments. The technique is non-
contacting and is suitable for on-line inspection, measurement of thickness of surface coatings,
and changes in metallurgical condition.
2-14
ET is most widely applied to non-ferromagnetic materials, although continuing advances have

adapted the method to ferromagnetic materials. Perhaps the main use in power plant components
is the in-situ examination of tubing in condensers and heat exchangers. Tubular inspection is
accomplished by either passing the tube through an encircling coil or by passing a
circumferentially wound bobbin through the tube. For the later inspection, a differential coil and
multi-frequencies are typically used. Portions of the tube that have been wasted by corrosion can
be located, the extent of metal loss quantified, and the deterioration periodically tracked with
successive examinations.
Portable Alloy Analysis
Material confirmation is an essential element in any failure analysis. Complete confirmation

would validate all specified mechanical properties (including hardness), fabrication history
(forming and heat treatment), metallurgical characterization, and chemical composition. Of
these, validation of the chemical composition has come to be known as Positive Material
Identification (PMI) and the advent of portable instruments permit measurement of a significant
number of elements using portable instruments in the field. PMI is not a substitute for the more
accurate laboratory data to be obtained later in the failure investigation but can provide an early
indication of whether the proper materials were used. Additionally, PMI can be used routinely
during condition assessments to validate materials and weld metals.
The primary basis for PMI is application of either portable x-ray fluorescence (XRF) instruments
or optical emission spectrometers (OES). The state-of-the art with these instrument is
continually improving, but XRF instruments typically are incapable of quantitative
measurements for elements with an atomic number lower than 22 (titanium and below). In
particular, the XRF method is incapable of measuring carbon (atomic number 6) which is a
critical element in many of the alloys for which PMI is desired. For carbon steels, the
application of portable optical emission spectrometers (OES) is appropriate to measure carbon
and other elements (e.g., silicon) not readily detected by XRF. Thus, PMI does not typically
provide absolute confirmation of all mandatory elements in the materials specification in the
same sense as the more sophisticated laboratory methods such as wet chemistry, optical emission
spectroscopy, energy dispersive spectroscopy, combustion/infrared spectroscopy, atomic
absorption spectroscopy, etc. Nevertheless, when coupled with other documentation, such as a
certified mill or material test report (CMTR), a Certificate of Compliance (COC), or material
marking, PMI can establish a high degree of confidence that the material matches what was
intended.
The XRF instruments are available with either an x-ray tube or a radioactive isotope. In
instruments with a radioactive isotope, the inherent source decay requires that the source be
periodically changed. Additionally, there are strict regulations controlling handling and traveling
with radioactive isotopes. These complications don’t exist with instruments employing an x-ray
tube because radiation is only present when the instrument is energized. The XRF instruments
are highly portable and are completely nondestructive. On the other hand, the OES instruments
are bulky and generate an arc strike on the component being tested to create the optical spectrum.
In some situations it may be necessary to remove the metal affected by the arc strike.
2-15
Ferrite Measurement
Ferrite measurement techniques evolved after the realization that austenitic stainless steel weld
metals, containing a moderate amount of ferrite, were free of hot cracking related weld defects.
Ferrite measurement was immediately identified as a method by which engineers could quantify
the amount of weld metal ferrite and ensure that their fabrications would be free from hot
cracking [2-10].
The advent of duplex stainless steels, with higher levels of ferrite than that for austenitic weld
metals, further re-emphasized the need for adequate ferrite measurement techniques as a suitable
ferrite/austenite phase balance provides adequate mechanical properties and improved corrosion
performance [2-10].
There are four approaches to the measurement of ferrite:

• Metallographic point counting
• Constitution diagrams
• Magnetic attraction
• Magnetic permeability
Of these, metallographic point counting is difficult to effectively implement and is seldom used
except to calibrate and validate the other methods.
Constitution diagrams are based on the concept that the phases are determined by ferrite
stabilizers and austenitic stabilizers. The diagrams are represented graphically as Nickel
Equivalent on the ordinate and Chromium Equivalent on the abscissa with the austenitic, ferritic,
and martensitic phases delineated. The four common constitution diagrams are Schaeffler,
DeLong, WRC 1988, and WRC 1992 [2-11, 2-12, 2-13, 2-14]. Beginning with the Schaeffler
diagram in 1956 and continuing through the WRC 1992 diagram, each change represents an
attempt to address a deficiency in the previous diagrams. As a generalization, the WRC 1988 is
preferred for “300” stainless steel weld metals. Figure 2-2 shows the Delong diagram (circa
1956) and Figure 2-3 shows the WRC 1988 diagram [2-15].
There are several instruments which use a magnet with a known strength and measure the
attractive force between the test piece and the instrument. The most common is a Magne Gage
which brings the reference magnet in contact with the test piece and measures the force required
to separate the magnet from the test piece (tear-off force). The Magne Gage is calibrated to
convert the tear off force to the ferrite content.
2-16
Figure 2-2
DeLong (FN) Diagram for Estimating the Ferrite in Stainless Steel Weld Metal [2-15]
2-17
Figure 2-3
WRC 1988 (FN) Diagram for Estimating the Ferrite in Stainless Steel Weld Metal [2-15]
Magnetic permeability is defined as the ratio of magnetic induction to magnetic strength. Ferrite
measurement using this technique requires that a magnetic field be induced into the test piece
and the resulting field strength be measured to establish magnetic permeability. Experimentation
established that the strength of the induced field varied with the amount of ferromagnetic phase
present. The Fisher Feritscope® was developed as a hand-held device which utilized the
magnetic permeability as a method to assess ferrite content. The Feritscope® was designed to be
portable and provide the operator with a user-friendly interface which readily provides ferrite
content on the ferrite number (FN) scale. Of the options available to measure ferrite content, the
Feritscope® is by far the easiest to use.
2.4 Preliminary Examination of Failed Component
Much of the discussion in this section was adapted from Reference 2-16.
The failed part, including all its fragments, should be subjected to a thorough visual examination
before any cleaning is undertaken. Often, soils and debris found on the part provide useful
evidence in establishing the cause of failure or in determining a sequence of events leading to the
failure. For example, traces of paint found on a portion of a fracture surface may provide
evidence that a crack, into which some paint seeped, was present in the surface for some time
before complete fracture occurred. Such evidence should be recorded.
2-18
The preliminary examination should begin with unaided visual inspection. The unaided eye has
exceptional depth of focus, and the ability to examine large areas rapidly and to detect changes
of color and texture. Some of these advantages are lost when any optical or electron-optical
device is used. Particular attention should be given to the surfaces of fractures and to the paths
of cracks. Initiation sites for cracks are especially important as well as crack path relative to the
loading on the part. It should be noted whether the crack path is straight or jagged and whether
crack branching occurs.
The significance of any indications of abnormal conditions or abuse in service should be

observed and assessed, and a general assessment of the basic design and workmanship of the part
should also be made. All important features, including dimensions, should be recorded, either in
writing or by sketches or photographs. It cannot be emphasized too strongly that the
examination should be performed as inquisitively and effectively as possible, because clues to
the cause of breakdown often are present but may be missed if the observer is not vigilant. A
low-power microscope (about 6 to 25 diameters), preferably of a binocular type, is an invaluable
aid in detection.
Where fractures are involved, the next step in preliminary examination should be general
photography of the entire fractured part, including broken pieces, to record their size and
condition and to show the fracture is related to other components. This should be followed by
careful examination of the fracture. The examination should begin with the use of direct lighting
and proceed at various angles of oblique lighting and dark-field illumination, to delineate and
emphasize fracture characteristics. This should also assist in determining which areas of the
fracture are of prime interest and which magnifications will be possible (for a given picture size)
to bring out fine details. When this evaluation has been completed, it is appropriate to proceed
with photography of the fracture, recording what each photograph shows, its magnification, and
how it relates to the other photographs.
Next, the nondestructive tests covered in the prior section should again be reviewed. Some NDE
may have already been performed but it is appropriate at this time to take a fresh look at prior
testing and to plan for additional NDE pending removal of samples.
2.5 Sample Removal, Identification, and Preservation of Essential Features

Such as Fracture Surfaces and Deposits
The essential considerations in removal of the failed component for further laboratory
examination are: (a) the preservation of all aspects of the sample that may be critical to the
failure analysis, (b) minimal introduction of solid or liquid contaminants which may later cloud
the interpretation, and (c) protection of the remaining field assembly from introduction of debris
and contaminants. The achievement of these three goals is situation specific and requires the
exercise of judgment and should involve the concurrence of field personnel who operate and
maintain the equipment.
Sample identification, either beforehand or immediately following sample removal, is essential.

This can be achieved by marking directly on the sample or by photography to preserve
2-19
orientation and such features as steam flow and flue gas directions, up/down/left/right relative to
some coordinate convention, critical dimensions, and adjacent components.
The method for removing the sample depends on the size, the material, and the
repair/replacement strategy. Thermal cutting (carbon arc electrodes, oxyacetylene, or plasma)
may be appropriate in some circumstances but adequate distance should be provided to avoid
introducing artifacts into the critical portion of the sample. Since thermal cutting processes
produce slags and spatter, the effect these may have on subsequent examinations should be
weighed.
Least disturbance of the sample is usually achieved with dry cutting processes using electric or
air powered reciprocating or circular saws or rotary grinders with thin cut-off disks (ziz-wheels).
If absolutely necessary, cutting fluids may be used with the recognition that these may prejudice
portions of the later examinations.
If there are external deposits which are loosely adherence, some should be collected in
polyethylene bags for later examination. If there are separated fracture surfaces, the urge to
“stick them back together” should be suppressed since this will damage certain features of the
fracture. The fracture surfaces should be protected by covering with bubble wrap of rags and
taping to the component. To preserve adherent deposits and scales, severe mechanical shocks
should be avoided during sample removal, packaging, and shipment. It is generally good
practice to “bag” each sample in a polyethylene wrapper using a water absorbing desiccant in the
bag and to further protect the sample with bubble wrap, newspapers, or rags for transport. The
protected sample should be shipped in a box (preferably wooden) surrounded with newspapers or
styrofoam popcorn to minimize movement and provide maximum protection against rough
handling.
Upon arrival at the laboratory performing the failure analysis, the failed component will be
additionally photographed for “as-received” documentation, assignment of laboratory tracking
codes, and divided into samples for metallographic examination, chemical analysis, and
mechanical testing.
Because the examination tools used by the laboratory, including hardness testers, and optical and
electron microscopes, are limited as to the size of the specimen they can accept, it is usually
necessary to remove multiple specimens for the laboratory examination. It is important that
records, either sketches or photographs, be kept to show the locations of the cuts made during
sectioning. The same precautions mentioned relative to sample removal apply equally to
sectioning.
2.5.1 Metallographic Samples
The metallographic samples will generally be allocated to those for visual and low-power
stereoscopic examination, those for optical microscopy, those for scanning electron microscopy
(SEM), and rarely, those for transmission electron microscopy (TEM). The sizes of the
specimens will be dictated by the particular instruments being utilized.
2-20
2.5.2 Chemical Analysis Samples
Some chemical analysis procedures consume the specimen (e.g., LECO carbon analysis), others
partially consume the specimen (e.g., optical emission spectroscopy) whereas others preserve the
specimen (e.g., x-ray fluorescence). In all instances there is a range of specimen size that is
suitable for the analytical procedure. Furthermore, to gain confidence in the results, it may be
desirable to repeat the analysis on multiple specimens which are thought to be “identical.” Close
coordination between the laboratory and the principal failure analyst is essential to plan the
chemical analysis portion of the investigation.
2.5.3 Mechanical Testing Samples
The type of mechanical tests and the extent of testing must be carefully planned upfront since
these tests tend to use a substantial amount of material. In some cases miniature samples can
yield nearly the same information as full size samples but requires more care in sample
preparation and sometimes requires different test apparatus. Sample orientation relative to the
component product (i.e., in or transverse to the principal working direction) effects some tests
more than others and needs to be planned carefully. Additionally, sample orientation relative to
a major fracture should be given proper consideration.
2.6 Metallographic Examination
To gain the best understanding of the failure, the metallographic examination should be done
over a range of magnifications, beginning with visual (1X) and increasing the magnification
incrementally to the highest level appropriate for the situation. Thus, a range of examinations
from the macroscopic (~1X to 50X) to the microscopic (50X to 1500X for optical light
microscopic and up to 20,000X for SEM) may be appropriate. A range of lighting conditions
should be utilized.
2.6.1 Macroscopic
Macroscopic examinations generally cover the magnification from 1X (visual examination) to

about 50X and are accomplished either visually (1X), with low magnification optically, or with
low magnification in a scanning electron microscope (SEM).
2.6.1.1 Visual, Low Power Microscopic, SEM
The examinations above 1X magnifications are best accomplished with a stereo microscope that
allows ease of zooming and provides the depth perception only possible with a three-dimensional
view. The part can be manipulated under the microscope as different lighting is applied. Many
of the features that are apparent when viewed in the stereo microscope are not nearly so readily
photographed. As magnification is increased, the extent of area that can be maintained in focus
2-21
is decreased. Photography may be easier when viewed in the scanning electron microscope
(SEM) at low magnifications.
Low power SEM examinations often are performed after the initial stereo microscopic
examinations. The SEM has the advantage of a large depth of field which allows substantial
fracture surface area to be maintained in focus.
The amount of information that can be obtained from low-magnification examination of a

fracture surface is extensive. Consideration of the configuration of the fracture surfaces may give
an indication of the stress system that produced the failure (pure tension, tension plus bending,
torsion, tension plus torsion, biaxial, etc.) The fracture surface also gives insight into the mode
of fracture (fatigue, brittle, ductile, mixed mode).
Macroscopic examination can usually determine the direction of crack growth and hence the
origin(s) of the failure. With brittle flat surfaces, determination depends largely on the surface
showing “chevron marks” which indicate the origin and direction of travel of the fracture.
Where fracture surfaces show both flat (implying brittleness or high constraint) and slant
(implying ductile) fractures, it may be generally concluded that the flat fracture occurred first.
Cracks may initiate and grow initially under high constrain (plane strain) as flat fractures. Crack
extension, often with crack front tunneling, relaxes the plane strain state of stress so that final
fracture occurs by slant fracture under plane-stress conditions. Conversely, if a fracture has
begun at a free surface, the fracture-origin area is usually characterized by a total absence of
slant fracture or “shear lip” [2-16].
If corrosion processes are involved, macroscopic examination can give insight into whether a
single process is operative or whether transitions occur in the modes of attack. Low-power
examination of fracture surfaces often reveals regions having a texture different from the region
of final fracture: fatigue, stress-corrosion and hydrogen embrittlement features may all show
these differences [2-16].
2.6.1.2 Hardness
Hardness testing is the simplest of the mechanical tests and is one of the most versatile tools
available to the failure analyst. Among its many applications, hardness testing can be used to:
(a) assist in evaluating heat treatment (comparing the hardness of the failed component with that
prescribed by specification), (b) provide an approximation of the tensile strength of steel, and (c)
detect work hardening, or to detect softening or hardening caused by overheating, by
decarburization, or by carbon or nitrogen pickup. Hardness testing is also essentially
nondestructive [2-16].
For macroscopic examination, the hardness tests to be utilized should sample a reasonable
volume of “affected material” encompassing several grains to properly relate to the bulk
hardness. The two most common types of hardness tests relevant to metals used in power plants
are static indentation tests and dynamic hardness tests.
2-22
In static indentation tests, a ball, diamond come, or pyramid is forced into the material being
tested. The relationship of total test force to the area or depth of indentation provides the
measure of hardness. The Rockwell, Brinell, Knoop, and Vickers hardness tests are of this type
and are the most commonly applied bulk hardness tests [2-17].
In dynamic hardness tests an indenter of standard mass and dimensions is bounced from the
workpiece; its height of rebound becomes a measure of hardness. This rebound (Schleroscope)
hardness testing method was developed by Albert F. Shore in 1907 and the hardness measure is
sometimes referred to as the Shore Schleroscope hardness. A similar method propels an indenter
into the workpiece and measures the approach and rebound velocities to calculate hardness on
the Leeb scale. This method of measurement is popularly employed in small hand-held portable
instruments [2-17].
Hardness should always be reported in the “native units” for the measuring instrument being
utilized and conversions to other scales should be noted. All such conversions between scales
are approximations and the converted hardness depends, in part, on the materials and hardness
levels.
2.6.1.3 Polishing and Etching
For macroscopic viewing, it is not necessary to achieve a high polish on a surface and a flapper
wheel with a 300 grit paper is usually sufficient. The as-polished surface typically reveals few
features of interest and macroetching is advised to assure no critical features are missed.
Samuels defines macroetching as the etching of a metal surface with the objective of
accentuating the gross structural details for observation by the unaided eye or at magnifications
not exceeding ten diameters [2-18]. A macrograph is a graphic reproduction of a prepared
surface of a specimen at a magnification not exceeding ten diameters. When photographed, the
reproduction is known as a photomacrograph. The etchants used in macroetching are generally
aggressive since the intent is to show gross detail rather than fine detail. Appendix A contains a
compilation of etchants used for macroscopic and microscopic sample preparation. The features
that are typically sought when macroetching are solidification patterns, working patterns, welds
and heat-affected zones, surface coatings, alloy segregation, and general alloy classification
judged by the response to the macroetch.
2.6.1.4 Interpretation
The main purpose of the macroscopic examination is to develop the “big picture” associated with
the failure analysis. It is perhaps the most essential initial step since the groundwork for more
detailed investigations begins with a thorough understanding of the results from the macroscopic
examination. Macroscopic examination of the failed component should assist in:
• Establishing the fracture origin
• Establishing the crack propagation path
• Identifying the mode of fracture (ductile or brittle)
2-23
• Determining whether there is a weld and the extent of involvement of the weld in the failure
• Identifying the role of environment (carburized or decarburized surface)
• Establishing the role of corrosion, and whether the corrosion attack is general or localized
• Measuring hardness, giving a partial basis to judge the relative strength
• Assessing strength in relation to the material specification
2.6.2 Microscopic
As with hardness testing and macroscopic examination previously discussed, the examination of
metallographic sections with a microscope is standard practice in most failure analyses, and is
often the mainstay in developing and documenting an understanding of the failure. Both optical
light microscopes and scanning electron microscopes may be used for the microscopic
examination, and on occasion the transmission electron microscope. Optical light microscopy
(OLM) remains the major method employed because of convenience of the preparation route,
relatively low costs, the intuitiveness of use, and familiarity of interpretation. The dominance of
OLM in failure analysis accrues to the outstanding capability of the microscope in revealing
material imperfections caused during processing and in detecting the results of a variety of in-
service operating conditions and environments that may have contributed to failure. Inclusions,
microstructural segregation, decarburization, carbon pickup, improper heat treatment,
untempered "white" martensite, and intergranular corrosion are among the many metallurgical
imperfections and undesirable conditions that can be detected and analyzed by microscopic
examination of metallographic sections [2-17].
Even in the absence of a specific metallurgical imperfection, examination of metallographic

sections is invaluable to the investigator in the measurement of parameters, such as case depth,
thickness of plated coatings, grain size, and heat-affected zone -- all of which may have a bearing
on the cause of failure. Metallographic sections are also useful when quantitative metallographic
techniques, such as point counting, lineal analysis, or electron-probe microanalysis, are
employed in failure analysis [2-17].
2.6.2.1 Polishing and Etching
Preparation of a metallographic specimen suitable for microscopic examination is, in part, an art
form that requires good dexterity, visual acuity, and attention to details. The sizes of specimens
which are to be viewed with an optical microscope are necessarily small, typically less than 0.2
square inches, to facilitate the preparation and viewing.
Unless the specimen is fairly large, it is mounted, usually in a thermoset plastic such as a
phenolic. Many other techniques may also be used but the basic requirements are that: (a) the
mounting material should be chemically inert with respect to the metal specimen, (b) the
mounting temperatures should not affect the structure, and (c) there should be no crevices
between the specimen and the mounting material [2-19].
2-24
Grinding is performed to remove flowed cut surface metal and to attain a flat surface. This must
be a cutting action, not buffing, which flows metal under pressure from high to low spots.
Grinding must be done so as to avoid any heating; therefore, water is used to cool and to carry
away cutting debris. Successively finer abrasive grits, each used at right angles to its
predecessor, permit cutting to proceed with smaller and smaller residual grooves.
Polishing is still a cutting operation, rather than a buffing or flow operation. For decades,
metallographic polishing has been performed on rotating wheels covered with a long-napped
cloth in which a very fine abrasive, usually alumina or a fine diamond paste, is suspended. With
the correct degree of pressure, a bright reflective surface with minimal metal flow can be
attained. In addition to conventional mechanical polishing just described, two new techniques,
vibratory polishing and electropolishing, are also widely used [2-20].
Vibratory polishing is popular because of the extremely high quality of the scratch-free polished
surface, the large number of specimens which can be prepared simultaneously, and the
possibility of freeing trained technicians from tedious polishing to devote their time to more
gainful matters.
Electropolishing is another process that has been gaining increasing acceptance as a means of
metallographic specimen preparation. The advantages of electropolishing over mechanical
polishing are: (a) rapidity with which specimens may be polished, (b) elimination of cold-worked
surface, (c) flatness of polished area, (d) application to a wide variety of materials, and (e)
polishing and etching can often be accomplished in one operation.
Polished metal specimens usually show no structural characteristics. However, on occasion

useful information, such as the presence of cavities or porosity, can be gained by microscopically
viewing a polished surface. Application of darkfield or differential interference contrast (DIC)
illumination facilitates imaging of features on polished surfaces. However, it is usually
necessary to etch the surface to render a complete interpretation for brightfield illumination.
Etching, or the use of a chemical which will differentially attack or dissolve metal of differing
chemical reactivity, is the final and most critical step of metallography. Etching of the metal
surface is done to make visible the crystalline structure or the metal and to produce optical
contrast between the various constituents [2-20]. Etching is done by immersing the cleaned and
polished specimen in a suitable etching solution such as one of those described in the
compilation in Appendix A.
Etchants are composed of organic or inorganic acids, alkali’s, or other complex substances in
some solvent such as water, alcohol, glycerin, or; glycol. These etching reagents are powerful
and must be handled with care. Since each reagent has been developed for a specific purpose it
must be chosen carefully if it is to reveal the structure that is desired. For example, picral is not a
general-purpose etchant but rather is designed to distinguish ferrite and iron carbide by
preferential darkening of the carbide phase. Nital, on the other hand, is not suited for the above,
but is used principally as a general etchant for steel and for delineating grain boundaries in
ferrite.
2-25
The specimen is usually held by tongs and immersed with the polished face down into a small
petri dish partly filled with the reagent. The specimen may also be swabbed with cotton which
has been saturated with the etchant. The progress of etching can be observed visually but should
also be timed. The proper etching time must be found experimentally and may vary from a few
seconds to a minute or more [2-20].
If the specimen is not sufficiently etched after the first immersion, the process may be repeated.
If the specimen is over etched, it must be re-polished prior to re-etching. Immediately following
the etching the specimen should be washed in water to stop the etching, then immersed in
alcohol and finally dried in a blast of warm air. Rapid drying is important to prevent water spots.
One of the purposes of chemical etching is to remove deformed metal which may have
developed during polishing. Etching reveals structural characteristics by preferential etching.
That means that some areas such as grain boundaries are more highly stressed than other areas
and are more subject to chemical etching. Also, the etching rate is different for different
crystallographic planes which results in varying degrees of reflected light with consequent
shading of the various grains.
2.6.2.2 Microhardness
The most common microhardness testing uses the fixed load indentation method and is
performed at low loads relative to the more conventional hardness testing covered in Section
2.6.1.2. The test loads vary from 1 to 1000 gram force (gf) and the load must be reported in the
test results.
The two most common microhardness tests are Vickers and Knoop. In both methods the
hardness value is determined by measuring the size of the permanent indentation by using a
microscope and established formulas or look-up tables in accordance with ASTM E384 [2-21].
The Vickers test dates to about 1925 whereas the Knoop test originated in 1939.
The Vickers microhardness indenter is geometrically the same as that used for the higher load
Vickers test. The indenter is a 136° diamond pyramid which produces a square indentation in an
isotropic material. Variable loads are employed dependent on the situation, enabling the use of
one hardness scale for all ranges of hardness – from very soft lead to tungsten carbide [2-17].
With the Vickers indenter, the depth of indentation is about one-seventh of the diagonal length.
The hardness is reported as a Vickers hardness number, HV. Historically it was also reported as
the Diamond Pyramid Hardness number, DPH, a term still occasionally used. Thus,
microhardness reported as HV or DPH is the same.
The Knoop indenter is a highly polished, rhombic–based pyramidal diamond that produces a
diamond –shaped indentation with a ratio between the long and short diagonals of about 7 to 1
[2-17]. The pyramid shape used has an included angle of 172° 30” and an included transverse
angle of 136°. The depth of indentation is about 1/30th of its length. The hardness is reported as
the Knoop hardness number, HK. As with the Vickers test, a variable load is used for the Knoop
test, dependent on the test situation. For a given load, the Knoop indenter penetrates to about
one-half the depth of the Vickers indenter [2-17].
2-26
Figure 2-4 compares indentations in the same material under loads of 3000, 1000, 500, and 100
gf [2-17]. The Knoop is often favored near an edge or in narrow zones, with the long diagonal
axis being parallel to the edge or the narrow zone.
Figure 2-4
Indentations made by Knoop and Vickers Indenters in the Same Work Metal at the Same
Load [2-17]
Specific applications for microhardness testing in production testing and failure analysis include
[2-17]:
• Measuring hardness of precision work- pieces that are too small to be measured by
conventional macroscopic hardness testing methods
• Measuring hardness of product forms such as foil or wire that are too thin or too small in
diameter to be measured by conventional macroscopic methods
• Monitoring of carburizing or nitriding operations, which is usually accomplished by hardness
surveys taken on cross sections of test pieces that accompanied the work pieces through
production operations
• Measuring hardness of individual micro- constituents
• Measuring hardness close to edges, thus detecting undesirable surface conditions such as
grinding burn and decarburization
• Measuring hardness of surface layers such as plating or bonded layers
• Measuring hardness gradients in heterogeneous regions such as the heat-affected zone (HAZ)
of weldments
2.6.2.3 Interpretation
The interpretation of microstructure and hardness from the microscopic examination requires a
comprehensive understanding of the physical metallurgy of the alloy involved. As a beginning
point, it is always advantageous to identify the procurement specification and to examine the
2-27
degree of conformance of the failed component to the requirements of the specification. It is

also essential to understand that the fabricator likely performed additional operations on the
procured product (tube, pipe, plate, sheet, bar, forging, and casting) such as bending, swaging,
flaring, welding, coating, and heat treatment, that caused changes in some properties relative to
those specified in the procurement document. To the extent possible, comparisons with newly
fabricated components, or to similar components which have not failed, are beneficial.
Some of the fundamental things to be learned from a microscopic examination are as follows:
• Grain size and whether equiaxed or directionally distorted
• Defects and whether they were present in the original product, created during fabrication, or
were a result of service
• Phases (for iron-based alloys, ferrite, pearlite, bainite, martensite, austenite, etc.)
• Carbides and their distribution
• Secondary phases such as sigma, Laves, etc.
• Nonmetallic inclusions (type, size, shape, location – grain boundary or matrix)
• Voids (cavities), their location, and origin
• Morphology of cracks (intergranular, transgranular, multiple or single origin)
• Hardness relative to expectation from procurement specification
• Uniformity of hardness
There are many compilations of microstructures and fractographs that will assist in the
interpretations from the microscopic examination [2-1, 2-2, 2-3, 2-18, 2-22, 2-23]. A wide
assortment of textbooks and seminar compilations can provide useful insights. Finally, personal
contacts in the industry with engineers who have encountered similar failures are highly
beneficial.
2.7 Chemical Analysis
In any failure investigation, it is strongly recommended that routine chemical analysis be

performed to ensure that the material conforms to that which was intended. Slight deviations
from specified compositions of many intended alloying elements or unintended residual elements
are not likely to be of major importance in causing failures. A background in physical
metallurgy will aid in ascertaining the significance of deviations. Attributing a failure to slight
deviations from the intended composition is seldom the correct interpretation.
A bulk chemical analysis is the usual basis to judge whether the correct material has been used.
However, even if the bulk chemical composition is as expected, there may be local
heterogeneities that are related to the failure which can be ascertained with other types of
laboratory instruments.
2-28
2.7.1 Bulk
There are multiple choices of methods and equipment to perform bulk chemical analysis
including wet chemical analysis, optical emission spectroscopy, atomic absorption spectroscopy,
x-ray fluorescence spectroscopy, and combustion methods.
Wet chemical analysis is the oldest technology but has waned in use as reliable and accurate
instruments have evolved to perform analysis based on spectrographic and combustion methods.
The two basic methods used in wet chemical analysis are gravimetry in which a chemical species
is determined by weighing and titrimetry which involves volume measurement of liquid
reactants. Wet chemistry is still used to validate calibrations and as a referee method to resolve
disputes in results. Wet chemical analysis methods are employed for determining the presence
- - --
and concentration of anions such as Cl , NO3 , and S . These methods are very sensitive [2-16].
Optical emission spectroscopy is perhaps the most useful analytical method because multiple
elements can be analyzed in a single arc burn. In the typical instrument the sample is placed in a
burn chamber and a pre-purge with argon is followed by a short period or arc burn to get stable
performance, followed by a longer period of arc burn in which the optical spectrum is sampled to
gather the information for numerical processing. All modern instruments are interfaced to
computers to perform the numerical processing. The entire process takes about 20 seconds and
can provide results for about 30 elements per burn, depending on the number of channels in the
spectrograph. The minimum sample size is about 0.5 inch (12 mm) diameter by 0.02 inches (0.5
mm) thick. The arc consumes some of the sample. Thus, over time the standards used for
calibration are consumed and have to be replenished since periodic calibration with known
samples is an essential aspect of instrument use.
There are many variants of atomic absorption (AA) spectroscopy including flame atomic
absorption, inductively coupled plasma (ICP), direct coupled plasma (DC) and graphite furnace
atomic absorption (GFAA). The method is mainly used for elemental detection at very low
concentrations, often for deleterious residual elements, in the parts per million (ppm) and sub-
ppm levels. In traditional flame AA, the sample is dissolved in a reagent which is then slowly
aspirated into a flame and the resulting spectrum is compared to a reference lamp containing the
element being analyzed. By changing reference lamps, more than one element can be measured
from an adequate amount of dissolved solution.
X-ray fluorescence (XRF) is the emission of characteristic “secondary” (or fluorescent) x-rays
from a material that has been excited by bombarding with high energy x-rays or gamma rays.
The characteristic wavelengths associated with each element are detected with a detector set at
the appropriate angle associated with the wavelength. The fluorescent radiation may be analyzed
either by sorting the energies of the photons (energy dispersive analysis) or by separating the
wavelengths of the radiation. Once sorted, the intensity of each characteristic radiation is
directly related to the amount of each element in the material. This is the basis of a powerful
technique in analytical chemistry. In the XRF wavelength dispersive method (WDX or WDS)
being discussed here, the intensity is graphed against the wavelength and the spectral peaks are
identified with specific elements. With multiple elements the characteristic peaks may have
some overlap and the numerical analysis corrects for these inter-element effects. A typical
2-29
wavelength scan moves the analyzing crystal through a large angular extent to permit
identification of a large number of elements. Sample size is comparable to that discussed for
OES is required. However, the sample is not consumed or altered by the analysis process and
calibration samples may be used repeatedly, an advantage of XRF over OES.
Carbon is routinely analyzed by a combustion method in a carbon determinator, often

manufactured by a company named LECO. This method is also used for sulfur, nitrogen,
hydrogen, and oxygen analysis. All methods involve exposure of the sample, usually in the form
of machined chips, in a high temperature furnace to react the elements of interest. For carbon
and sulfur an infrared detection (IR) system is employed and this same approach is used for
nitrogen, hydrogen, and oxygen with the addition of the measurement of thermal conductivity
(TC). The IR and TC measurements correlate to the levels of the indicated element in the
sample.
2.7.2 Local
The most common method to measure local micro-constitutients in a sample is by the energy
dispersive analysis of x-rays (EDAX) in a scanning electron microscope (SEM). The region of
interest can be located using normal viewing, such as backscattering of x-rays, and the EDAX
feature used to gather data on chemical composition. The same result can be achieved with a
wavelength dispersion analysis of x-rays (WDX) attachment to a SEM but the process is slower
than EDAX and less commonly employed. A SEM with an electron microprobe is utilized to
examine small areas for chemical composition. The Auger electron spectrometer is sometimes
applied to the analysis of thin surface layers [2-16].
Compound identification is usually accomplished with x-ray diffraction. Each crystalline

compound has a characteristic x-ray diffraction signature. Electron diffraction using a thin foil
in a TEM or an extraction replica can also be used for compound identification in crystalline
materials.
2.8 Mechanical Testing
Mechanical tests are sometimes used to demonstrate that features (fracture type, fracture path,
ductility, etc.) associated with the failure are similar to the same features from a standard
laboratory test, or a non-standard test devised to simulate the failure. The most commonly
applied laboratory tests used in a failure investigation include:
• Smooth bar tensile test
• Notched tensile test
• Smooth bend test
• Notched bend test
• Charpy V-notch impact test or other relevant fracture toughness tests such a KIc or JIc
• Creep or stress rupture
2-30
• Fatigue
Tensile test are often employed to demonstrate that the yield and tensile strength are as expected
and the ductility (elongation and reduction in area) is consistent with expectations. The service
exposure (duration, temperature, and environment) may have caused some change in these
properties so comparison to the procurement specification may not be appropriate for a failed
component. Additionally, all the change may not be due to service exposure but may have come
from the initial fabrication methods employed to make the component. Thus, some judgment
must be exercised in interpreting the tensile test results. If the failure displays low ductility in
proximity to a stress riser, a notched tensile test may be useful. Similar information can be
gained from a smooth and notched bend test. If a calibrated machine test isn’t possible, a
hammer-blow with sufficient force to cause a fracture can provide insights into the basic material
behavior. While room temperature tensile tests are usually sufficient, in some instances it may
be desirable to perform the test nearer to the temperature at which the component failed.
It is important to characterize the component failure as predominantly ductile or predominantly

brittle. If brittle fracture is involved, some Charpy-V-notch tests are useful to establish the
ductile-to-brittle transition temperature (DBTT) and to understand if this played a role in the
failure. If more quantitative information is necessary, plane strain fracture mechanics tests may
be performed to establish KIc, or elastic-plastic fracture tests employed to establish JIc, from
which KIc can be inferred.
Creep or stress rupture tests (creep implying the use of extensometers to measure creep strain
whereas stress rupture tests mainly measure the time to failure under constant load and
temperature) may be appropriate for failures associated with elevated temperature operation.
These tests are inherently of long duration and careful consideration should be given to the test
conditions.
If the failure mode is fatigue, it may be beneficial to perform fatigue tests. Room temperature
fatigue tests are readily performed and there are usually published curves for comparison, but
fatigue tests at high temperature are substantially more difficult, as is the comparison to expected
behavior.
2.8.1 Options Based on Material Availability and Support to Failure Type
From the previous section, it is obvious that the available material, the budget, and the time will
all play a role in the type of mechanical testing chosen to support the failure analysis. Standard
sample sizes are preferable because they ease the interpretation of results. However, sub-size
sample can be employed on some tests to conserve material. Tensile test results are not usually
sensitive to specimen size and there are standard adjustments of test values to account for sub-
sized Charpy V-notch specimens. If specimen size prevents the measurement of plane strain
fracture toughness, and elastic-plastic JIc test can be performed on a smaller specimen and KIc
inferred from the results. Another alternative for fracture testing is the small punch test for
assessing the fracture appearance transition temperature, FATT [2-3]. For creep and stress
rupture testing, the role of oxidation must be considered in selecting specimen size. Testing in
2-31
an inert environment is an available option to negate the oxidation limitation, but few facilities
offer this type of testing.
2.8.2 Interpretation
The basis for interpretation critically depends on specific details of the failure. As an initial
basis, the mechanical test results should be compared to the procurement specification. If there
is adequate knowledge to adjust for changes that occurred in fabrication, these should be
included. If embrittlement played a role, the embrittling effect may be predictable [2-3]. For
Charpy V-notch tests, there are a number of correlations that convert measured energy to plane
strain fracture toughness [2-3].
The expectation from the mechanical tests is that they provide insights into the failure. It may be
possible that the initial tests are not definitive and additional tests emerge as necessary to provide
useful information. These are all judgment issues and require experience to reach reasonable
conclusions.
Sometimes it is beneficial to perform finite element stress analysis of the failed component. In
that situation various scenarios can be analyzed reflecting the spread in the measured mechanical
properties to place the mechanical test results in perspective.
2.9 Identification of Failure Mechanism(s) Based on Collective Information

Gathered
The diversity of equipment and materials in a fossil power plant give rise to a large number of
failure mechanisms that either occur alone or in combination with other failure mechanisms.
Most design codes only provide formal protection from a limited subset of the full complement
of failure mechanisms and seldom include rules to safeguard against failures in which
environment plays a major role.
While component-specific failure mechanisms are too large to enumerate in this section, the
three volumes by Dooley and McNaughton on boiler tube failures are noteworthy because more
than 30 failure mechanisms have been identified [2-26 – 2-28]. A partial compilation of the
main failure mechanisms for boiler tubes is shown in Table 2-3.
The remainder of this section contains an abbreviated list of potential failure mechanisms that are
of concern in fossil power plant components and identification of the failure mechanism for a
specific instance may not be included here. Thus, the analyst should search other sources for
additional failure mechanisms, especially those that are component specific [2-1, 2-2, 2-3, 2-6,
2-16, 2-22, 2-23, 2-24].
2-32
Table 2-3
A Partial Compilation of Failure Mechanisms for Boiler Tubes [2-26 – 2-28]
Water-Touched Tubes
ID Sutface Initiation OD Surface Initiation
Corrosion Fatigue Flyash Erosion
Hydrogen Damage Waterwall Fireside Corrosion
Acid Phosphate Corrosion Supercritical Waterwall Cracking
Caustic Gouging Sootblower Erosion
Erosion-Corrosion in Economizer
Inlet Tubes & Headers Coal Particle Erosion
Chemical Cleaning Damage: Acid-Dewpoint Corrosion:
Waterwalls Economizer
Pitting
Steam-Touched Tubes
ID Sutface Initiation OD Surface Initiation
Pitting SH/RH Sootblower Erosion
Stress-Corrosion Cracking Long-Term Overheating
SH/RH Fireside Corrosion: Coal-
SH/RH Chemical Cleaning Damage Fired Units
SH/RH Fireside Corrosion: Oil-
Fired Units
Rubbing/Fretting
Stress-Corrosion Cracking
Gross Distortion or Overload
Properly designed and operated equipment is not expected to encounter this failure mechanism,
especially early in life. Later in life when corrosion mechanisms have reduced the available
cross-section (sound metal) to resist load, this failure mechanism becomes more probable.
Anomalous operational events such as overheating or overpressure should be pursued as
potential root causes.
Whether the failed component displays substantial distortion or negligible distortion is an

important consideration. If there is substantial distortion it is apparent that the material displayed
ductility and measurements may quantify the associated strains. If there is limited distortion the
material could either be brittle or ductile but with a triaxial stress state producing sufficient
constraint to mask the display of macro ductility. In this latter event, the fracture surface will
2-33
still show features of ductile fracture such as microvoid coalescence. Constraint is promoted by
high stress concentration factors and abrupt section changes.
Brittle failure can be attributed to materials that inherently have low ductility, such as grey cast
iron, to embrittlement [2-3], or to operation in the proximity of the lower shelf for body-centered
cubic materials such as ferrous alloys which display a ductile-to-brittle transition temperature.
Creep or Stress Rupture
As a matter of convenience the mechanical behavior of materials is divided into a low

temperature region where the strength properties are essentially independent of time and strain
rate and a high temperature region where the properties are time dependent. In the low
temperature region, the sciences of elasticity and plasticity are typically applied to describe the
stresses and strains in a component. As temperature increases the strength properties become
increasingly dependent on time and deformation continues even at the same load. The time-
dependent, thermally assisted deformation of components under load (stress) is known as creep.
If the loads persist until failure, the failure mode is called stress rupture. Laboratory tests in
which only the failure time is measured are called stress-rupture tests, whereas tests in which the
deformation is measured as time progresses, but in which the tests are terminated prior to failure,
are called creep tests. Likewise, laboratory tests in which the creep strains are measured and are
permitted to continue to failure are called creep-rupture tests.
The temperature at which creep becomes a consideration for the materials used in fossil power
plant components is in the range of about 0.3 to 0.5 of the absolute melting temperature. The
temperatures where creep begins to govern the allowable stresses are in the range of 700-800F
(371-427C) for carbon steels, 900-950F (482-510C) for low-alloy Cr-Mo steels, and 1000-1150F
(538-621C) for the austenitic stainless steels.
The creep rates expected with long term service dictate that the failures will be intergranular and
the damage typically begins with voids in the grain boundary associated with non-metallic
inclusions, carbides, or at triple points. Early in life these cavities are isolated but with the
passage of time the grain boundaries aligned perpendicular to the principal tensile stress,
gradually displaying aligned cavitation. With additional passage of time, microcracks join the
cavities and this progression continues until macrocracks develop and may ultimately lead to
through-wall failure.
The above sequence may not occur with short-term overheating and the failure sequence will be
more like a tensile overload.
The macrostrains associated with long term creep-rupture failures are typically below 5 percent,
much smaller than observed in a hot tensile test. This lower strain, and the knowledge that the
operating temperature is in the creep range, aids in classifying the failure mechanism.
Metallographically the intergranular fracture path, multiple cracks, and the presence of creep
voids provide additional evidence.
Fatigue
2-34
It has been recognized since 1850 that a metal subjected to a repetitive or fluctuating stress will
fail at a stress much lower than required to cause fracture on a single application of load.
Failures occurring under conditions of dynamic loading are called fatigue failures, presumably
because it is generally observed that failures occur after a considerable period of service [2-25].
There is no obvious change in structure of a metal which has failed in fatigue which can serve as
a clue to an understanding of the reasons for fatigue failures, at least at usual magnifications
associated with failure analyses.
With few exceptions, fatigue is associated with the presence of stress risers which provide higher
local stresses. Once fatigue cracks originate, their progression is dominated by the stress field,
thus they tend to propagate transgranularily and nominally form planar cracks.
The cyclic stress may be caused by cyclic mechanical loads (start-stop cycles on rotating
equipment), by flow induced vibrations, and by fluid temperature changes on either surface of a
component. The fluid temperature changes may occur on the fluid side of heat absorbing tubes
or on the furnace side from variable heat flux, slag accumulation, slag shedding, and soot
blowing. The cyclic stresses arising from thermal sources are associated with thermal fatigue.
The more sophisticated design codes (ASME III, VIII-2 and VIII-3) using the design-by-analysis
approach formally consider fatigue in the design stage whereas many of the simpler design-by-
rule codes (ASME I) do not. In essence, these simpler codes typically provide an “implicit”
fatigue design margin from their conservative component sizing.
Creep-Fatigue
Many components which operate at sufficiently high temperatures to incur creep deformation
also experience concurrent cyclic loading. If both tensile and compressive creep stresses occur
on a component in service, it becomes obvious that there are an infinite number of “path
dependent” cyclic possibilities. The usual outcome of operation with creep fatigue is that a more
rapid accumulation of damage occurs, thus life is shorter than when either mechanism operates
alone. If either creep or fatigue dominant, the usual failure features associated with the dominant
mechanism will be displayed. However, without a dominant mechanism, it is difficult to
generalize on the macro deformation and microstructural damage mechanisms. The “damaging
synergism” should be suspected if the failure cannot be explained as purely creep or purely
fatigue.
Corrosion-Fatigue
Corrosion fatigue occurs by the combined actions of cyclic loading and a corrosive environment.
It is a particularly prevalent in fossil boilers where corrosion fatigue occurs frequently on the
water side of economizer tubes and headers, waterwall tubes and headers, risers, downcomers
and drums, with a preference toward regions with increased local stresses. While the
mechanisms of crack initiation and growth are complex and not fully understood, there is
consensus that the two major factors are strain and waterside environment. Strain excursions of
sufficient magnitude to fracture the protective oxide layer play a major role. In terms of the
waterside environment, high levels of dissolved oxygen and pH excursions are known to be
2-35
detrimental. Historically, the steels applied in these water-touched components have had the
minimum specified yield strengths in the range of 27 to 45 ksi (185 to 310 MPa) and minimum
specified tensile strengths in the range of 47 to 80 ksi (325 to 550 M Pa). As these materials are
supplanted by higher strength steels, there is concern that the higher design stresses and thinner
wall thicknesses will render components more vulnerable to failures by corrosion fatigue. Thus,
when employing such higher strength steels for water circuits in boilers it is desirable to use
"best practices" in: (a) design by minimizing localized strain concentrations, (b)in control of
water chemistry and during layup by limiting dissolved oxygen and pH excursions, and (c) in
operation by conservative startup, shutdown, and turndown practices.
Microscopically, corrosion fatigue cracks are characterized by features such as multiple, wide,
transgranular cracks with irregular profiles usually filled with oxide and blunt tipped, and
showing signs of discontinuous growth such as crack arrest marks. In Reference 2-27, Chapter
13 has 41 pages covering this failure mechanism in boilers.
General Corrosion
General corrosion, also called uniform attack, is the most common form of corrosion. It is
normally characterized by a chemical or electrochemical reaction which proceeds uniformly over
the entire exposed surface or over a large area [2-29]. In the limit, the metal becomes thinner
and fails by overload.
In fossil plant components the loss of metal due to high temperature oxidation and fireside
wastage are two of the most common problems in boilers. This type of wastage has become
more acute with the introduction of low NOx firing systems which impede the formation and
growth of stable oxides which provide protection to the surface.
Localized Corrosion
There are many categories of localized corrosion. Those of major importance to fossil power
plant components include crevice corrosion, pitting, intergranular corrosion, stress corrosion, and
hydrogen damage. These will not be discussed in detail presently because of the expansiveness
of the subject but the interested reader may review appropriate references [2-26 – 2-32].
2.9.1 Results are Definitive and Failure Mechanism Conclusive
Similar to the legal system, most failure analyses are based on demonstrating that the associated
failure mechanism is supported by the “preponderance of evidence.” Some of the evidence may
have minor departures from that associated with the failure mechanism but the cost and expense
of additional investigation must be balanced against the benefit to be gained. When there is high
confidence that the results are sufficiently definitive to support the failure mechanism, the formal
investigation is essentially complete and the analyst can proceed to Section 2.10.
2-36
2.9.2 Results are Inconclusive and Additional Investigation Required
In the “first pass” of a failure investigation it may be ascertained that the results are inconclusive
and that additional investigation is required. In such an event, the principal failure analyst
should review the sequential steps previously outlined in this Chapter to pinpoint the additional
information that is needed. It will probably be beneficial to re-convene a team from the
disciplines and organizations identified in Figure 2-1 to establish the future direction and get
agreement on funding and schedule. If any disciplines or organizations were not included in the
initial round, consideration should be given to broadening the team to include their participation
for the additional investigation. If there is a division of opinion on the interpretation of results
within the team, it is often beneficial to engage “new blood,” either within or outside the
organization, to get a different perspective.
2.10 Synthesis of Results from Failure Analysis and Unit Operation to

Establish a Root Cause
The failure mechanism and root cause are linked together but they are distinctly different
concepts. The root cause is the “driver” which activates the failure mechanism. Thus, for each
failure mechanism there is a possible multiplicity of root causes. On occasion there is a singular
root cause which drives the failure but it is not uncommon to identify more than one root cause.
The root cause is typically associated with the following factors

• Design
• Fabrication and Erection
• Operation
• Materials
• Environment
• Maintenance
Components are typically designed to an industry standard but occasional they are designed to a
company’s proprietary standard. Component design typically requires knowledge of materials,
geometry, loads, pressures, temperatures, and environments and it usually provides a margin to
account for uncertainties in the fundamental design parameters. If one or more of the design
parameters, coupled with the uncertainty, errs on the non-conservative side, the probability of
failure is heightened. To the maximum extent possible, the failure analysis should establish the
component design basis in order to establish whether it is inadequate and becomes a root cause.
The fabrication and erection of a component usually involves forming loads and strains far in
excess of those encountered in service and thermal cycles, including welding and heat treatment,
which alter the material properties. Construction Codes have some provisions to control
fabrication practices but they may not account for all situations. An understanding of the details
2-37
of fabrication and erection for each component will assist the analyst in determining whether
they are root causes.
Operation relates to the path followed by all the relevant parameters to start the plant, operate it
through the full range of load, shut the plant down, and lay it up to mitigate out-of-service failure
mechanisms, mainly corrosion phenomena. The scenario also includes occasional emergency
shutdowns and overspeed events for rotating equipment. Data logging equipment is employed in
modern plants to log the essential data and facilitate retrieval of historical practices. By trending
key data it can be established when certain aspects of operation depart from the norm in a
manner which may identify a potential root cause. Personal interviews with key plant personnel
involved in operations may provide insights into root causes associated with operation.
Materials, including base metals, weld metals, coatings, and claddings, are likely to play a
pivotal role in failure analyses and may sometimes be the root cause. Application of the wrong
material may be innocuous in some cases or may pave the way for premature failure in others.
Surface protection in the form of coatings or claddings may eventually be consumed in service
and establish a root cause. Use of the wrong material that has a deficit in strength, ductility,
corrosion resistance, or heat resistance relative to the intended material may establish the root
cause. But if the wrong material does not have such associated deficits, it is not valid to ascribe
the root cause to the use of the wrong material. Material impairment or embrittlement from
service operation, e.g., graphitization, sensitization, sigma phase formation, and creep cavitation,
can be the root cause for a failure.
Environment plays a key role in many failures. The role of environment must be viewed over
the entire operational cycle for a particular component from pre-operation through startup, on-
load operation, shutdown, and layup. Aggressive detrimental effects may be associated with one
or more of the events in the operational cycle. For example, water side corrosion fatigue in
boiler components is usually associated with the pre-operational, startup, shutdown, and layup
periods but seldom with on-load operation. For water/steam side damage, the basic water
treatment protocols (all volatile treatment, the various phosphate treatments, oxygen treatment,
caustic treatment) and dates of major change are important to understanding failure in the boiler
and turbine [2-26 to 2-28]. On the flue gas side, the design-basis fuel as well as a history of fuel
switches can be useful in unraveling trends in failure, especially boiler tubes. Additionally,
changes in firing systems, including switches to low NOx firing and use of over-fired air, are
important in understanding boiler tube failures. Finally, metal wastage associated with
environmental effects gradually erodes the ability of the component to resist the applied loads
and can be a root cause.
Maintenance of power plant equipment typically involves three potential downside issues; (a) it
isn’t performed correctly, (b) it creates an unintended and undesirable consequence as a result of
the maintenance, or (c) it isn’t performed soon enough and damage is already done. All of these
possibilities should be evaluated as potential root causes of failures. In the first two cases, the
failure often occurs shortly after the maintenance activity and can be identified as the root cause.
In the latter instance, it may be difficult to established delayed-maintenance as the root cause, or
the association may be apparent, e.g., loss of bearing lubrication due to unrepaired oil-line leaks.
2-38
2.11 Assemble and Organize the Evidence, Formulate Conclusions, Write

Report, Including Options to Avoid Repeat Failures
Sadly, this step is often seen as the least fulfilling one because the excitement of the “chase” is
over and engineers usually view writing as “drudgery.” But it is a highly crucial step to preserve
a record for future reference. Laxity in doing the best job in writing the report will eventually be
regretted by those who have a follow-up role on a repeat or similar failure.
To assure that all necessary information has been collected before writing the report, Vander
Voort poses a comprehensive series of questions that are worthy of repetition here [2-33].
• Has failure sequence been established?
• If failure involved cracking or fracture, have the initiation sites been determined?
• Did cracks initiate at the surface or below the surface?
• Was cracking associated with a stress concentrator?
• How long was the crack present?
• What was the intensity of the load?
• What was the type of loading: static, cyclic, or intermittent?
• How were the stresses oriented?
• What was the failure mechanism?
• What was the approximate service temperature at the time of failure?
• Did temperature contribute to failure?
• Did wear contribute to failure?
• Did corrosion contribute to failure? What type of corrosion?
• Was the proper material used? Is a better material required?
• Was the cross section adequate for class of service?
• Was the quality of the material acceptable in accordance with specification?
• Were the mechanical properties of the material acceptable in accordance with specification?
• Was the component that failed properly heat treated?
• Was the component that failed properly fabricated?
• Was the component properly assembled or installed?
• Was the component repaired during service and, if so, was the repair correctly performed?
• Was the component properly run in?
• Was the component properly maintained? Properly lubricated?
2-39
• Was failure related to abuse in service?

• Can the design of the component be improved to prevent similar failures?
• Are failures likely to occur in similar components now in service, and what can be done to
prevent their failure?
The answers to these questions should have mostly been derived from a combination of records
and the examinations and tests previously discussed in this Chapter.
The failure-analysis report should be written clearly, concisely and logically. In the absence of a
Company standard, a suggested outline for the major sections of the report follows.
1. Abstract
2. Description of the failed component
3. Service conditions at time of failure
4. Prior service history
5. Manufacturing and processing history of component
6. Mechanical and metallurgical study of failure
7. Metallurgical evaluation of quality
8. Summary of failure mechanisms and root cause analysis
9. Recommendations for prevention of similar failures or for correction of similar components

presently in service.
Obviously, not every report will require coverage under every one of these sections. The
language in the report should be appropriate to the audience. If communication to non-technical
organizations is an important consideration, the minimization of technical jargon is highly
desirable. The use of appendixes, containing detailed calculations, equations, and tables of
chemical and metallurgical data, can serve to keep the body of the report clear and uncluttered.
2.12 References
2-1. Boiler Tube Failure Metallurgical Guide, Volume 1: Technical Report. EPRI, Palo Alto,
CA: 1993. TR-102433-V1.
2-2. Boiler Tube Failure Metallurgical Guide, Volume2: Appendices. EPRI, Palo Alto, CA:
1993. TR-102433-V2.
2-3. Embrittlement of Components in Fossil Fueled Power Plants, EPRI, Palo Alto, CA:
2003. 1004515.
2-40
2-4. Guideline on Fossil Boiler Field Welding, EPRI, Palo Alto, CA: 2003. 1004701.
2-5. C. R. Weymuller, “How to Use NDT to Find Defects,” Welding Design and Fabrication,
(October 1990).
2-6. Condition Assessment Guidelines for Fossil Fuel Power Plant Components, EPRI, Palo
Alto, CA: 1990. GS-6724.
2-7. Fossil-Fired Boiler Tube Inspection, Volume 1: Nondestructive Testing Guidelines,

EPRI, Palo Alto, CA: 1986. CS4633.
2-8. W. R. McDearman, “Practical Aspects of the Field Inspection of Boiler Tubes,” in

Failures and Inspections of Fossil-Fired Boiler Tubes: 1983 Conference and Workshop,
EPRI, Palo Alto, CA: 1983. CS-3272.
2-9. Metals Handbook, Ninth Edition – Volume 17: Nondestructive Evaluation and Quality
Control, ASM International, Metals Park, OH: 1989.
2-10. C. D. Lundin, W. Ruprecht, and G. Zhou, Literature Review: Ferrite Measurement in

Austenitic and Duplex Stainless Steel Castings, Materials Joining Research Group,
Department of Materials Science and Engineering, The University of Tennessee,
Knoxville, August 1999.
2-11. A. L. Schaeffler, “Constitution Diagram for Stainless Steel Weld Metal,” Metals Progress
56(11): 680-680B, 1949.
2-12. W. DeLong, G. Ostrom, and E. Szumachowski, “Measurement and Calculation of Ferrite

in Stainless Steel Weld Metal,” Welding Journal 35 (11), 521-s to 528-s, 1956.
2-13. T. A. Siewert, C. N. McCowan, and D. L. Olson, “Ferrite Number Predictions to 100FN

in Stainless Steel Weld Metal,” Welding Journal 67 (12):289-s to 298-s, 1988.
2-14. D. J. Kotechi and T. A. Siewert, “WRC-1992 Constitution Diagram for Stainless Steel
Weld Metals: A Modification of the WRC 1988 Diagram,” Welding Journal, Vol. 71.
171-s to 172-s, May 1992.
2-15. ASME Specification SFA-5.4, Specification for Stainless Steel Electrodes for Shielded
Metal Arc Welding, Section II, Part C, Specifications for Welding Rods, Electrodes, and
Filler Metals, American Society of Mechanical Engineers, New York, NY 2004.
2-16. Metals Handbook, Eighth Edition, Vol. 10: Failure Analysis and Prevention, American
Society for Metals, Metals Park, OH, 1975.
2-17. Metals Handbook, Ninth Edition, Vol. 8: Mechanical Testing, American Society for
Metals, Metals Park, OH, 1985.
2-41
2-18. L. E. Samuels, Optical Microscopy of Carbon Steels, American Society for Metals,
Metals Park, OH, 1980.
2-19. Robert M. Brick, Alan W. Pense, and Robert B. Gordon, Structure and Properties of
Engineering Materials: Fourth Edition, McGraw-Hill Book Company, New York, NY,
1977.
2-20. Dell K. Allen, Metallurgy Theory and Practice, American Technological Society,
Chicago, IL, 1977.
2-21. “Standard Test Method for Microhardness of Material,” E 384, Annual Book of ASTM
Standards, ASTM, Philadelphia, PA.
2-22. ASM Handbook Volume 9: Metallography and Microstructures, ASM International,

2-23. ASM Handbook Volume 12: Fractography, ASM International, Metals Park, OH, 1987.
2-24. Charlie R. Brooks and Ashok Choudhury, Metallurgical Failure Analysis, McGraw-Hill,
Inc., New York, 1993.
2-25. George E. Dieter, Mechanical Metallurgy: Second Edition, McGraw-Hill Book

Company, New York, 1976.
2-26. R. B. Dooley and W. P. McNaughton, Boiler Tube Failures: Theory and Practice,
Volume 1, Boiler Tube Fundamentals. EPRI, Palo Alto, CA: 1995. TR-105261-V1.
Volume 2, Water-Touched Tubes. EPRI, Palo Alto, CA: 1995. TR-105261-V2.
Volume 3, Steam-Touched Tubes. EPRI, Palo Alto, CA: 1995. TR-105261-V3.
2-29. Mars G. Fontana and Norbert D. Green, Corrosion Engineering: Second Edition,
McGraw-Hill Book Company, New York, NY. 1978.
2-30. ASM Handbook Volume 13A: Corrosion: Fundaments, Testing, and Protection, ASM
International, Metals Park, OH, 2003.
2-31. ASM Handbook Volume 13B: Corrosion: Materials, ASM International, Metals Park,
OH, 2005.
2-32. ASM Handbook Volume 13C: Corrosion: Environments and Industries, ASM
International, Metals Park, OH, 2006.
2-33. G. F. Vander Voort, “Conducting the Failure Examination,” Met Eng Quart, May 1975.
2-42
3
TYPICAL LABORATORY METHODS AND EQUIPMENT
UTILIZED
3.1 Imaging
Microscopy is the science of the constitution and structure of metals and alloys and at the light
optical level is concerned with the proportions, distributions, and dispersion of phases in
microstructures [3-1]. Microstructure is the structure of a prepared surface of a metal or alloy
revealed by a microscope when enlarged to magnifications greater than about 10X. Imaging by
direct viewing with a microscope, photography, or display on a video monitor is a key element in
interpreting microstructures.
3.1.1 Visual and Low Power Stereo Microscopy
As discussed in Section 2.6.1, most failure analyses begin with low magnification imaging
(macroscopic) ranging from 1X (visual examination) to about 50X using low power stereo
microscopes. In the magnification range from 10X to 50X there is a transition from macroscopic
examination to microscopic examination. The general object of the low power examination is to
understand the overall features of the failure
A stereo microscope that allows ease of zooming and provides the depth perception only possible
with a three-dimensional view is the preferable means for performing the low power
examination. The part can be manipulated under the microscope as different lighting is applied.
Many of the features that are apparent when viewed in the stereo microscope are not nearly so
readily photographed. As magnification is increased, the extent of area that can be maintained in
focus is decreased. Photography may be easier when viewed in the scanning electron
microscope (SEM) at low magnifications because of the large depth of field offered by the SEM.
The stereo zoom microscope is usually mounted on a boom arm which facilitates moving the
component being examined underneath the objective lens of the microscope as illustrated in
Figure 3-1. Illumination can be provided by transmitting light through bendable fiber optic
cables or by a ring fluorescent lamp surrounding the objective. The flexibility to view features
of interest through a wide range of magnifications using the zoom is a valuable feature of these
microscopes. Zoom rations on the order of 15:1 are available which could nominally cover a
magnification from 3X to 45X. Photographic attachments are available on many stereo zoom
microscopes.
The role of the stereo zoom microscope and the useful date that can be gathered in a failure
analysis was discussed in Section 2.6.1 and will not be repeated here.
3-1
Typical Laboratory Methods and Equipment Utilized
Figure 3-1. Illustration of a Stereo Zoom Microscope Mounted on a Boom Arm as Typically
Used for Low Power Three-Dimensional Viewing of Failed Components
3.1.2 Optical Light Microscopy (OLM)
Optical light microscopy (OLM) refers to the observation of structures that can be resolved by
means of an optical microscope [3-1]. Examination of metals by use of the optical microscope
requires that a section be cut, that the surface of the section be prepared to a high reflection
condition, and usually that this surface be etched suitably to develop the structure of interest.
In OLM, the microscope must accomplish three tasks: produce a magnified image of the
specimen (magnification), separate the details in the image (resolution), and render the details
visible to the eye, camera, or other imaging device (contrast). A magnifying glass is a simple
microscope. These magnifiers have severe limitations in specimen positioning, illumination,
lens aberrations, and construction. The highest practical magnification achievable with a hand
held magnifying glass is 8X to 10X and possibly 25X when mounted. This magnification is far
less than that needed for serious metallography.
For higher magnification, compound microscopes have been developed, achieving near-
theoretical performance levels. In its basic form, the compound microscope consists of two
convex lenses aligned in series: an object glass (objective) closer to the object or specimen, and
an eyepiece (ocular) closer to the observer’s eye—with means of adjusting the position of the
specimen and the microscope lenses. The compound microscope achieves a two-stage
magnification. The objective projects a magnified image into the body tube of the microscope
and the eyepiece further magnifies the image projected by the objective.
Compound microscopes are of two types; transmitted light and reflected light. Figure 3-1
illustrates the basic layout for a transmitted light microscope [3-2]. In transmitted light
microscopes, the light passes through the sample, hence the sample has to essentially be
3-2
Figure 3-2
Diagrammatic illustration of a Transmitted Light Compound Microscope in Which the
Intermediate Image formed by the Objective and Tube Lens is Enlarged by the Eyepiece
[3-2]
transparent. These are the types of microscopes mostly used in biological and medical
institutions. Essentially, a transmitted light microscope functions as follows. Light from a lamp
(not shown, but near the bottom of Figure 3-2) passes through a condenser containing an
adjustable iris and then through a transparent specimen placed over an opening in the stage.
Light is then gathered by the objective. The objective, together with the built-in tube lens,
focuses the image of the specimen at the level of the fixed diaphragm of the eyepiece. The
image is then seen by the observer as if it were at a distance of approximately10 inches (250
millimeters) from the eye [3-2].
When you look into a microscope, you are not looking at the specimen; you are looking at an
image of the specimen. The image is “floating” in space about 10 millimeters below the top of
the observation tube (at the level of the fixed diaphragm of the eyepiece) where the eyepiece is
inserted. The image you observe is not tangible; it cannot be grasped. It is a “map” or
representation of the specimen in various colors and/or shades of gray from black to white. The
expectation is that the image will be an accurate representation of the specimen, accurate as to
detail, shape and color/intensity. The implications are that it may well be possible (and is) to
produce (or even enhance) highly accurate images [3-2].
3-3
For metals or other opaque specimens, a reflected light microscope is utilized. The performance
is very similar to the transmitted light microscope except the light arriving at the specimen is
reflected back through the objective lens, the built-in tube lens, and the image is similarly
focused at the level of the fixed diaphragm of the eyepiece to appear as if it were at a distance of
approximately 10 inches (250 millimeters) from the eye. The reflected light microscope is the
most commonly used tool for the study of the microstructure of metals. Figure 3-3 shows a
photograph of a typical metallurgical (reflected light) compound microscope [3-3] and Figure 3-
4 illustrates the lens and light path used for these instruments [3-4].
Figure 3-3
Photograph of a Typical Bench-Type Reflected Light Compound Microscope Used in
Metallography [3-3]
The main optical components of a microscope are its two imaging lenses (eyepiece and
objective) and a condenser lens. The eyepiece and objective are responsible for magnifying the
image of the specimen and projecting it onto the viewer’s retina or onto the film plane in a
camera. The job of the condenser lens is to focus a cone of incident light onto the specimen. To
provide the incident light there is an illumination system which may include the source of the
incident light or may direct external natural or artificial light towards the condenser lens. It can
also provide means for enhancing the contrast and detail seen in the image. Finally, there is a
movable stage which holds the specimen in the optical path and allows the specimen to be
moved in and out of the focal plane and even left, right and rotated about the optic axis.
3-4
Figure 3-4
Illustration of the Reflected Light Microscope and the Trace of Rays through the Optical
Lens System [3-4]
To complete the instrument the microscope may include other attachments: filters, polarizers,
cameras, a viewing screen, and even hot and cold stages.
Contrast between different regions when viewed in reflected can arise from variations in surface
topography and differences in reflectivity (e.g. of different phases, different grain orientations, or
boundary regions). These features are revealed by a series of specimen preparation techniques
3-5
that should be carried out with care to produce high quality and useful images. Reflected-light
microscopy is used for a range of materials, including metals, ceramics and composites.
A multiplicity of imaging techniques have evolved for the reflected light microscope, most
aimed at increasing the contrast to facilitate image interpretation. Briefly, these are brightfield,
darkfield, polarized light, phase contrast, and differential interference contrast (DIC).
Brightfield
The preponderance of OLM for metals utilizes brightfield illumination in which the incident
light strikes the specimen at 90 degrees to the surface as illustrated in Figure 3-4. The way in
which the specimen scatters light will determine how well it can be imaged in this way. Natural
color is often obscured. Highly reflective specimen such as polished metals, ceramics, and
microchips are well suited for brightfield illumination. If the essential attributes discussed
previously (magnification, resolution, and contrast) are achieved with brightfield, there is little
motivation to try alternate illumination methods. The deficiency which usually drives the
microscopist to pursue other methods is a desire to achieve greater contrast.
Darkfield
In darkfield illumination the light is routed to the specimen in an annular outer ring where it exits
the objective and strikes the sample at an oblique angle rather that at a right angle as in
brightfield. In broad terms, the effect is to reverse the white-black regions from the brightfield
image, but there are other more subtle contrasting effects. In reflected light, darkfield will show
the true color of the specimen better than brightfield/
Polarized Light
For reflected light microscopy, contrast enhancement using polarized light is used for certain
optically anisotropic materials where conventional etching and imaging techniques fail to reveal
sufficient detail. This reason for improvement in contrast is as follows. When light enters the
polarizer (a rotatable polarizing film on the incident light side of the specimen), only the light
whose “vibrations” are parallel to the polarizing plane of the polarizer can pass through it. When
this plane polarized light reflects off an optically anisotropic metal surface, the light is reflected
as two components which are rotated and plane-polarized at right angles to each other. The
strength of each component and the degree of rotation varies depending on the crystallographic
plane. Finally, when this light passes through the analyzer (a fixed polarizing film in the path of
the reflected light), those elements of the light which are aligned with the analyzer may pass. All
others are extinguished. If the analyzer and polarizer at set at 90 degrees to each other (cross-
polarization), then only the rotated components of the reflected light will pass. In this way one
can obtain varying intensities from different grains since the orientation of each grain will be
different, exposing different crystallographic planes at the surface. If the specimen is rotated the
brightness of each grain will change.
3-6
Phase Contrast
Phase contrast microscopy has found greatest application in transmitted light imaging of
transparent (mainly cellular and biological) objects. In phase contrast microscopy additional
contrast is obtained by retarding the undeflected and diffracted rays by differing amounts before
recombining them to form the final image. One disadvantage is an artifact which appears as
halos around sharply delineated edges.
When used in reflected light microscopy this technique reveals slight, as small as 0.005 microns,
changes in height of the specimen. But like the halo artifact seen in transparent specimens, bands
of light and dark regions are often observed, for instance, along an inclined section of the
specimen. Phase contrast microscopy requires special condenser and objective lenses. These are
usually marked “Phaco” on the lens housing.
Differential Interference Contrast
Differential interference contrast (DIC) microscopy is similar to, but more complicated, than the
phase contrast interference system. The Nomarski imaging technique to enhance relief features
is a DIC method familiar to many metallurgist.. The DIC technique begins by polarizing the
incident light and then passing it through a prism which splits the light and rotates the two parts
90 degrees so that they are cross-polarized. This light now passes through or is reflected by the
specimen, is recombined using a prism and is once again polarized.
The result is high-resolution images, brilliant colors and a striking three-dimensional illusion.
The mechanism responsible for the contrast and color is similar to that in phase contrast method.
In reflected light microscopy the contrast comes from slight differences in the distance between
the prism and various spots on the surface of the specimen.
DIC does not produce the halos seen in phase contrast microscopy, but it does produce a few of
its own artifacts. The appearance of three-dimensions is an illusion. This can be demonstrated by
looking at a micrograph of a fairly flat specimen (i.e., a polished and etched metal) and noting
which parts appear raised and which appear lowered and then rotate the micrograph 180 degrees
and notice that the relief has been reversed. The colors are another artifact which, while they can
be fabulous, are false.
Comparisons of Imaging Techniques
Figures 3-5 through 3-10 present a series of photomicrographs in which a comparison can be
made between the images of various materials as they appear in brightfield, phase contrast, and
DIC [3-5]. The direction of shear in the DIC photomicrographs is approximately from the lower
left-hand corner to the upper right. Generally speaking, the information content of the DIC
photomicrographs is greater than those of the brightfield and phase contrast, but there are
exceptions.
3-7
For example, in Figure 3-5 a close examination will show that there are details in the DIC which
are not readily apparent in brightfield, but there also are details easily visible in brightfield which
probably would be overlooked if the specimen were examined only in DIC. Specimen details in
phase contrast are largely masked by the diffraction effects. In Figure 3-6, the information
content of the brightfield photomicrograph is practically nonexistent and while the information
content of the phase contrast photomicrograph is significant, it in no way compares to that of the
DIC. It is noteworthy that Figure 3-6 was mechanically polished but not etched, leaving the
sample highly reflective and without significant features in the brightfield image. However the
phase contrast and DIC imaged the small height relief created by the mechanical polishing.
While the general appearances of the three photomicrographs in Figure 3-7 are different, their
information content appears to be very nearly the same. In Figures 3-8 and 3-9 there are details
visible in the phase contrast photomicrographs which are less easily discerned in the DIC. Close
examination of the photomicrographs in Figures 3-10 and 3-11 shows a general superiority in the
image detail of the DIC photomicrographs. In examining these photomicrographs, and all the
preceding ones, it should be remembered that only the specimen structure which lies
perpendicular to the direction of shear is enhanced in the DIC image. If the specimens were
rotated 90° and reexamined, some of the details visible in Figures 3-5 through 3-11 would
disappear while other details not visible in these photomicrographs would emerge.
Figure 3-5
Photomicrographs of Ferrite at 200X. The Specimen was Mechanically Polished and Deep
Etched [3-5]
3-8
Figure 3-6
Photomicrographs of Cast Alloy at 200X.The Specimen was Mechanically Polished but
was not Etched [3-5]
3-9
Figure 3-7
Photomicrographs of Cadmium at 100X. The Specimen was Chemically Polished [3-5]
3-10
Figure 3-8
Photomicrographs of High Manganese Cast Steel at 125X. The Specimen was
Mechanically Polished but was not Etched [3-5]
3-11
Figure 3-9
Photomicrographs of Beta Brass at 400X. The Specimen was Mechanically Polished and
Etched [3-5]
3-12
Figure 3-10
Photomicrographs of Zircaloy at 400X. The Specimen was Chemically Polished [3-5]
3-13
Figure 3-11
Photomicrographs of Zirconium at 400X. The Specimen was Chemically Polished [3-5]
3.1.2.1 Magnification, Numerical Aperture, Resolution, and Depth of Field
The total visual magnification of an optical microscope is the product of the magnification
produced by the objective lens and the eyepiece lens. Thus, with a 40X objective lens and a 10X
eyepiece, the total magnification is 400X. From this concept, it would appear possible to
magnify to any extent desired by proper choice of lens. While magnification can be increased
indefinitely in this manner, the additional information gained by increasing magnification
reaches a threshold and increasing magnification produces larger images but no additional detail
[3-6]. Magnification beyond this threshold is called empty magnification and further
magnification beyond this threshold is fruitless. The fundamental limit on magnification is
determined by the wavelength of light rather than the combination of magnifying lens. More
will be discussed regarding this concept under the topic of numerical aperture and resolution.
The factor that determines the amount of image magnification is the objective magnifying power,
which is predetermined during construction of the objective optical elements. Objectives
typically have magnifying powers that range from 1X to 100X, with the most common powers
being 4X, 5X, 10X, 20X, 40X, 50X, and 100X. An important feature of microscope objectives is
3-14
their very short focal lengths. The primary reason that microscopes are so efficient at
magnification is the two-stage enlargement that is achieved over such a short optical path, due to
the short focal lengths of the optical components.
Eyepieces, like objectives, are classified in terms of their ability to magnify the intermediate
image produced by the objective lens. Their magnification factors vary between 5X and 30X
with the most commonly used eyepieces having a value of 10X to15X.
The ability to distinguish and discriminate between minute details lying close together in the
specimen is known as resolving power and is controlled by the numerical aperture (NA) of the
objective lens. The ability of an objective to include the various rays of light coming from each
illuminated part of the specimen is directly related to the angular aperture of the objective.
Objectives with lower angular aperture can include only a narrower cone of light as compared to
objectives with higher angular aperture [3-2]. The equation for numerical aperture is:
NA = n Sinα Eq. 3-1
In this equation:
NA is the numerical aperture
n is the index of refraction of the material in the space between the specimen
and the lens of the objective.
α is1/2 the angular aperture of the objective
For air, the index of refraction, n, is 1.00 and when oil immersion is used to couple the objective
lens to the specimen, the refractive index is about 1.51. The largest half-angle is about 72°,
making the largest value of sin about 0.95. Thus, in air the largest possible NA is 0.95 whereas
with oil coupling it is about 1.4. The value of the NA is affixed to each objective lens, usually
with air or oil immersion properly noted.
The important relationship between numerical aperture and resolving power will now be
examined. Resolving power has been defined as the ability of an objective to separate clearly
two points or details lying close together in the specimen. Resolution has been defined as the
actual (rather than theoretical) separation distance of two details lying close together still seen as
separate [3-2]. A separate definition of resolution relates to the ability to image a single point
feature. The fundamental equation for resolution is:
Cr λ
r= Eq. 3-2
NA
where:
r = resolution; for two minute points, the distance between the points lying close
together in the specimen but still showing the points as separate. For a single
point, the smallest point size which can be resolved, microns
Cr = the resolution constant. Values which are appropriate are:
3-15
= 0.25 for a single point [3-8]

= 0.50 for two points by the Sparrow criterion [3-8]
= 0.61 for two points by the Rayleigh criterion [3-8]
= the wavelength of light being used, microns
NA = the numerical aperture of the objective
From analysis of these equations, the following inferences can be made. As NA increases, r
becomes smaller; the size of the distance between adjacent points becomes smaller; hence
resolution is better. If shorter wavelengths of light are used (e.g. violet-blue end of the spectrum)
the resolvable distance becomes smaller; resolution is better. Longer wavelengths in the visible
light spectrum (e.g. red) yield poorer resolution. However, bear in mind that the human eye is
most sensitive in the green wavelength. Resolution varies inversely with numerical aperture.
Higher NA objectives are capable of yielding the best resolution; hence are better for separating
very minute details.
Visible light consists of that narrow portion of the electromagnetic spectrum to which the human
eye is sensitive. Wavelengths for visible light radiations lie in the range from 380 nm
(nanometer; 1000 nm = 1 μm) for radiation that provides a violet stimulus to the eye, up to 780
nm which provides a red stimulus to the eye. The ultraviolet spectrum is immediately below 380
nm and the infrared spectrum is immediately above 780 nm. If the eye is stimulated by light
containing all wavelengths of the visible region, the effect is that of white light; the sensation of
color results if wavelengths from one or more portions of the visible region are appreciably
absent. Figure 3-12 illustrates the wavelengths associated with the colors in the visible light
region [3-7].
Wavelength, nanometers
Figure 3-12
Portion of the Electromagnetic Spectrum Occupied by Visible Light [3-7]
A final equation is needed to complete the understanding between magnification (MP =

magnifying power) and resolution, r. As noted previously, the image viewed thorough the
eyepiece is at an effective distance of 250mm (10 in.) from the eye. At this distance, the visual
resolution for normal sight is 76 microns (0.003 in.). When the object is magnified by an optical
system, the visual resolution, R, at the object is thus:
3-16
76
R= Eq. 3-3
MP
If we now equate the visual resolution, R, from Eq. 3-3 to the resolution from Eq. 3-2, and solve
for the magnifying power, MP. The following result is obtained:
76 NA
MP = Eq. 3-4
Cr λ
where λ is in units of microns.

Table 3-1
Best Achievable Resolution and Associated Magnification for Optical Light Microscopy
Wavelength, λ Resolution, r Mag (Calc) Mag (Rule of Thumb)

Image Light
Medium Cr Feature NA microns Color microns Min. Max. Min. Max.
0.25 Single 0.145 525 1050
Air 0.50 Double 0.95 0.55 Green 0.289 263 525 475 950
0.61 Double 0.353 215 430
0.25 Single 0.118 642 1284

Air 0.50 Double 0.95 0.45 Blue 0.237 321 642 475 950
0.61 Double 0.289 263 526
0.25 Single 0.105 722 1444

Air 0.50 Double 0.95 0.40 Violet 0.211 361 722 475 950
0.61 Double 0.257 296 592
0.25 Single 0.098 774 1548

Oil 0.50 Double 1.4 0.55 Green 0.196 387 774 700 1400
0.61 Double 0.240 317 634
0.25 Single 0.080 946 1892

Oil 0.50 Double 1.4 0.45 Blue 0.161 473 946 700 1400
0.61 Double 0.196 388 775
0.25 Single 0.071 1064 2128

Oil 0.50 Double 1.4 0.40 Violet 0.143 532 1064 700 1400
0.61 Double 0.174 436 872
The phenomenon of diffraction and the limiting effect of the size of light waves dictate the “rule
of thumb,” that the useful magnification of an objective is 500 to 1000 times the numerical
aperture of the objective. (e.g., magnification range from 475 to 950 for an objective of 0.95
numerical aperture; magnification range from 700 to 1400 for an objective with a numerical
aperture of 1.4). Table 3-1 shows the best achievable resolution for air coupling (NA = 0.95) and
3-17
oil coupling (NA = 1.4) for green, blue, and violet light for the three values of the resolution
constant, Cr, previously discussed.
In Table 3-1, the minimum calculated magnification is based on Eq. 3-4 and the maximum
calculated magnification is arbitrarily set as twice the minimum. Many lens systems are
optimized to achieve the least distortion from chromatic and spherical aberration when using
green light (0.55 micron wavelength) and the other values in the table may not be achievable
with such lens. In general the detailed calculations validate the validity of the rule of thumb. For
OLM with air between the specimen and the objective lens using green light, these results
indicate that maximum magnifications of 215X to 1050X may be used and magnifications above
1050X are fruitless unless light with a shorter wavelengths is used. Likewise, with oil between
the specimen and the objective lens using green light, these results indicate that maximum
magnifications of 315X to 1550X may be used and magnifications above 1550X are fruitless
unless light with a shorter wavelengths is used. In the limit with short wavelength violet light,
the maximum useful magnification is 1440X with air coupling and 2130X for oil immersion.
Care should be taken in choosing eyepiece/objective combinations to ensure the optimal

magnification of specimen detail without adding unnecessary artifacts. For instance, to achieve a
magnification of 250X, the microscopist could choose a 25X eyepiece coupled to a 10X
objective. An alternative choice for the same magnification would be a 10X eyepiece with a
25X objective. Because the 25X objective has a higher numerical aperture (approximately 0.65)
than does the 10X objective (approximately 0.25), and considering that numerical aperture
values define an objective's resolution, it is clear that the latter choice would be the best.
A severe limitation to OLM is the inherent small depth of field. This practically negates use of
the instrument for fractographic examinations and is the fundamental reasons for the tedious
preparation route to guarantee a smooth and flat surface. Depth of field relates to the vertical
separation between the objective lens and the specimen. There is a “sweet spot” which is
achieved by moving the objective until a clear image is attained. Around this sweet spot there is
a range of objective lens movement which still achieves sufficient clarity for interpreting the
image. Depth of field is defined as the distance from the nearest object plane to the farthest
object plane which are in acceptable focus.
The depth of field for the objective lens can be expressed mathematically as [3-9]:
λ n 2 − NA 2
d = Eq. 3-5
NA2
where d is the depth of field in microns and the other variables are as previously defined
Typical values for depth of field in air are 8.5 microns (0.00033 in.) for low magnification
objectives with NA of 0.25 to 0.19 microns (0.0000075 in.) for high magnification objectives
with NA of 0.95.
3-18
3.1.2.2 Sample Preparation (mounting, polishing, etching)
This subject was covered in substantial detail in Sections 2.6.1.3 and 2.6.2.1 and there will be no
further elaboration on the subject in this section.
3.1.2.3 Phases and Particle Identification
Microscopy is principally a means to an end and not an end unto itself. The ultimate objective is
to understand how structures are formed, the factors that control their formation, and how they
determine significant physical or mechanical properties. Four steps are involved in attaining this
objective: (a) formation of the structure, (b) characterization of the structure, (c) measurement of
a property, and (d) establishment of a correlation between the structure and the property [3-1].
While there is not universal agreement, it can be argued that the four steps just outlined are the
foundations of physical metallurgy and mechanical metallurgy. The optical microscope is one of
the most import tools available to the metallurgist to characterize structure.
Several strengthening mechanisms exist for metals and alloys including substitutional solution
strengthening, interstitial solution strengthening, dispersion hardening, precipitation hardening,
multiphase strengthening, and strain hardening (cold work). To varying degrees, these
strengthening mechanisms impact structure in a manner that can be imaged by optical light
microscopy. But substantial supplementary information is necessary to provide the proper
interpretation of structure.
Equilibrium phase diagrams provide an essential complement to OLM to understand the physical
metallurgy. Binary phase diagrams are available for virtually all meaningful systems and ternary
phase diagrams for some common systems. Equilibrium phase diagrams may be classified
according to the relation of the components in the liquid and solid states as follows [3-10]:
1. Components completely soluble in the liquid state.

a. Completely soluble in the solid state.
b. Partly soluble in the solid state.
c. Insoluble in the solid state.
2. Components partially soluble in the liquid state.

a. Completely soluble in the solid state.
b. Partly soluble in the solid state.
3. Components completely insoluble in the liquid state.

a. Completely insoluble in the solid state.
The first type of equilibrium diagram in which the components are completely soluble in the
liquid state is most common in engineering alloys.
These equilibrium diagrams can be supplemented by isothermal time-temperature-

transformation (TTT) diagrams and continuous-cooling-transformation (CTT) diagrams to gain
3-19
a better understanding of the transformations that occur during cooling. These diagrams usually
show the expected hardness when cooling is complete. Jominy quench hardness tests showing
both the as-quenched hardness profile and the quenched plus tempered hardness profile provide
insights relative to bulk hardness over a range of cooling rates.
Samuel’s [3-1] provides a comprehensive summary of the optical microscopy of carbon steel and
similar documents exist for other alloying systems. By convention, single phases are identified
by Greek letters ( -ferrite, -brass, -solid solution, -phase, etc.) or chemical composition
(CuAl2, Fe3C, Mg2Si, TiN, etc.), and phase mixtures (eutectics, eutectoids, etc.) are described by
such terms as “lamellar,” “rod,” and “spheroidal” which allude to the prominent growth mode to
provide a general description of morphology.
The single exception to this kind of terminology for phase identification occurs in the
metallography of iron-carbon alloys as an almost exclusive evolution of the science from
observations made on steels and cast iron. For these iron-carbon alloys, names which are usually
associated with a pioneer researcher are used. The single-phase constitutients of iron-carbon
alloys, and by inference, most low alloy steels, from this evolution are:
• Austenite
• Ferrite
• Cementite
• Carbide
• Graphite
• Martensite
The names of the two-phase constitutients are:

• Pearlite
• Bainite
In OLM, phase and particle identification is mainly achieved by knowledge of the physical
metallurgy and comparisons of features of the structure with known (catalogued) similar alloys.
Hardness, including microhardness, can provide supportive information to identify phases and
particles.
3.1.2.4 Grain Size Estimates
Grain size is measured with a light microscope by counting the number of grains within a given
area, by determining the number of grains (or grain boundaries) that intersect a given length of
random line, or by comparing with standard grain-size charts. Microscopes fitted with an image
analysis system can automate the measurement. A very common method of measuring grain size
is to compare the grains at a fixed magnification with the American Society for Testing and
Materials (ASTM) grain size-charts as defined in ASTM Standard E19. The ASTM grain-size
number, n, is related to N*, the number of grains per square inch at a magnification of 100X, by
the relationship:
3-20
N * = 2 n−1 Eq. 3-6
Table 3-2 shows the relationship between n, N*, and the average grain diameter.
Table 3-2
Relationship between ASTM Grain Size Number and the Average Grain Diameter
n N* Average Grain Diameter

ASTM Grain Grains/in2
Size No. @ 100X in. mils mm microns
-3 0.0625 0.0394 39.4 1.000 1000
-2 0.125 0.0278 27.8 0.707 707
-1 0.25 0.0197 19.7 0.500 500
0 0.5 0.0139 13.9 0.354 354
1 1 0.00984 9.84 0.250 250
2 2 0.00696 6.96 0.177 177
3 4 0.00492 4.92 0.125 125
4 8 0.00348 3.48 0.0884 88.4
5 16 0.00246 2.46 0.0625 62.5
6 32 0.00174 1.74 0.0442 44.2
7 64 0.00123 1.23 0.0313 31.3
8 128 0.000870 0.870 0.0221 22.1
9 256 0.000615 0.615 0.0156 15.6
10 512 0.000435 0.435 0.0110 11.0
11 1024 0.000308 0.308 0.00781 7.81
12 2048 0.000217 0.217 0.00552 5.52
In ferrous alloys, the austenite grain size is extremely important in controlling the properties.
This influence is attributable to the effect of austenite grain size on the reaction characteristics
during transformation as the austenite is cooled. Table 3-3 is a simplified representation of the
relation between the grain size and properties [3-10].
Since fracture toughness is such an important property for steels, a study of the requirements for
pressure vessel quality plates in terms of grin size will be beneficial. All of the plate
specifications for power boilers reference American Society of Mechanical Engineers (ASME)
SA-20 [3-11] that contains the general requirements for plate that delineate the metallurgical
structure and melting practice comprising coarse grain (CG) and fine grain (FG) practice. When
a CG austenite grain size is specified, the steel has an American Society for Testing and
Materials (ASTM) grain size in the range of 1 to 5 (higher numbers are finer). The average grain
3-21
diameter for ASTM Number 1 is 0.0098 in. (0.25 mm) and for Number 5, 0.0025 in. (0.063 mm)
(see Table3-2).
Table 3-3
Simplified Summary of Grain-Size Effects on Properties of Steels
Austenite Grain Size

Property Fine Coarse
Depth of hardening Shallower Deeper
Retained austenite Less More
Warpage in Quenching Less More
Penetration in carburizing Shallower Deeper
Possibilitry of quenching cracks and
checking in grinding Less More
Possibilioty of soft spots in quenching More Less
Internal stresses after quenching Lower Higher
Machinability aftyer normalizing Inferior Better
Yield strength Higher Lower
Tensile ductility Higher Lower
Impact strength More Less
Fracture Toughness More Less
Ductile-to-brittle transition temperature Lower Higher
Creep strength Lower Higher
When a FG austenite grain size is specified, aluminum is usually used as the grain refining
element during melting, and the requirement is for an ASTM Number 5 or finer. FG practice
typically produces a grain size of ASTM Number 7 with an average grain diameter of 0.0012 in.
(0.031 mm). The aluminum content for FG practice should be in excess of 0.020% total, or
alternately, 0.015% acid soluble aluminum.
The effect of FG practice is to lower (improve) the ductile-to-brittle transition temperature,

giving the material higher fracture toughness, or stated differently, higher absorbed energy in a
Charpy V-notch test. This improvement in fracture toughness is especially important for the
hydro test, particularly when performed at 1.5 times design pressure. There have been a few
instances of drum failures during hydrostatic testing due to inadequate fracture toughness for the
test temperature employed.
The usual grain size for steels is from ASTM GS No. 1 to 9 with ASTM No. 1 to 5 representing
coarse grain steels and ASTM GS No. 5-9 representing fine grain steels.
3-22
3.1.3 Scanning Electron Microscopy (SEM)
3.1.3 Scanning Electron Microscopy (SEM)
The scanning electron microscope (SEM) has increasingly gained favor in the performance of
failure analyses. It offers four main advantages relative to optical light microscopy: (a) higher
magnification (b) resolution of smaller structures, (b) potential for measurement of elemental
composition with additional instrumentation, and (d) greater depth of field, which is of particular
benefit in fractography.
3.1.3.1 Magnification and Resolution Relative to OLM
High velocity electron beams that are produced when electrons are accelerated across large
voltages become wavelike in character. The effective wavelength for electrons used in electron
microscopes is on the order of 10-10 m [3-12]. Recall from the discussion of Figure 3-12 that the
wavelength of green light most commonly used in OLM is 0.55 X 10-6 m. Since resolution is
fundamentally related to the electromagnetic wavelength, the use of electrons accelerated across
a proper voltage field has the potential to improve resolution on the order of 5000 times as
compared to optical light microscopy, inferring magnification possibilities on the order of
5,000,000X. While the full extent of theoretical magnification is not realized in actual practice
with an SEM, magnifications to 100,000 X are achievable, an improvement over OLM of about
100. But more commonly, magnification in the range of 1000X to 10,000X is utilized.
3-23
Figure 3-13
A Diagrammatic Representation of a Scanning Electron Microscope [3-13]
The resolution of an SEM is basically controlled by the electron beam diameter rather than the
wavelength of an electron and is on the order of 0.0035 microns which is about 100 times
smaller than OLM using green light [3-13]. At the other end of the magnification scale,
fractographic examinations take advantage of the larger depth of field of the SEM and typically
utilize magnifications in the range of 10X to 1000X.
Figure 3-13 shows a diagrammatic representation of an SEM [3-13]. The electron beam is
collimated by the condenser lenses and then focused by the objective lens into a small-diameter
beam. This beam strikes the surface of the sample (the anode electrically), and the interaction of
the beam with the sample generates emitted electrons whose quantity is especially sensitive to
the surface topography. A scanning coil causes the beam to raster the surface of the specimen.
The quantity of electrons emitted from any point on the surface controls the intensity of a
synchronized video display, cathode ray tube (CRT) or otherwise. Hence, as the electron beam
rasters the surface, an image is generated on the video display which is essentially a picture of
the surface.
The interaction of the electron beam with the specimen surface generates several signals as
illustrated in Figure 3-14. The electrons emitted from the sample have varied energies. The
electrons with relatively low energies (secondary electrons, < 50eV) and those with high energies
(back-scattered electrons, 5 to 50 keV) are of particular interest. When the beam penetrates the
sample surface, the back-scattered electrons (BSE) are scattered elastically back out of the
sample, retaining approximately their original energy. Other electrons interact with the electrons
of the atoms of the specimen and have their energy changed. Most of these lose considerable
3-24
energy and are emitted with relatively low energy; these are the secondary electrons (SE). The
electrons penetrate the sample in a tear-drop-shaped volume, and only those within about 0.1
micron of the surface escape.
Figure 3-14
Illustration of the Emissions Produced by from the Sample in Response to the
Incident Electron Beam in the SEM [3-13]
The x-rays emitted from the sample surface are used for analysis of elemental composition using
either wavelength dispersion spectrometry (WDS) or energy dispersive spectrometry (EDS) [3-
13]. The EDS method is much faster than WDS but is does not detect light elements, such as
carbon, as well as WDS. These methods will be discussed more completely in Section 3.2.
3.1.3.2 Sample Preparation
Specimens examined in the SEM should be electrically conductive and must be connect
electrically as the anode to facilitate acting as a target for the electron beam. Virtually any shape
and size that can be contained in the vacuum chamber can be examined in the SEM. No special
sample preparation is required other than cleaning of the area to be examined. For polished and
etched metallography samples embedded in plastic, a conductive silver paint is applied to a
location on the sample away from the area of interest and electrically connected to the anode
circuit. This versatility of sample preparation is a major appeal of the SEM.
3-25
3.1.3.3 Viewing Modes
Imaging is typically obtained using secondary electrons for the best resolution of fine surface
topographical features. Alternatively, imaging with backscattered electrons gives contrast based
on atomic number to resolve microscopic composition variations, as well as topographical
information.
Secondary electron imaging provides high-resolution imaging of fine surface morphology. Inelastic
electron scattering caused by the interaction between the sample's electrons and the incident electron
beam results in the emission of low-energy electrons from near the sample's surface. The orientation of
surface features influences the number of electrons that reach the secondary electron detector which
creates variations in image contrast that represent the sample's surface topography. The secondary
electron image resolution for an ideal sample is about 0.0035 microns.
Backscattered electron imaging provides elemental composition variation, as well as surface topography.
Backscattered electrons are produced by the elastic interactions between the sample and the incident
electron beam. These high-energy electrons can escape from much deeper than secondary electrons, so
surface topography is not as accurately resolved. The efficiency of production of backscattered electrons
is proportional to the sample material's mean atomic number, which results in image contrast as a function
of composition - higher atomic number material appears brighter than low atomic number material. The
optimum image resolution for backscattered electron imaging is about 0.0055 microns.
3.1.3.3.1 Polished or Polished and Etched
As noted in the discussion of sample preparation, polished or polished and etched metallographic
samples can be examined so long as they are electrically connected to the anode portion of the
SEM to act as a target for the electron beam. There is limited reason to examine polished
samples from an imaging standpoint whereas polished and etched samples may be examined to
compare with OLM results, to extend to higher magnification, or to do elemental analyses.
3.1.3.3.2 Fracture Surfaces
Other than the SEM, fracture surface should be examined visually and at low magnification with
an optical stereo zoom microscope. But both of these approaches have only low magnification
capability and the images observed may be difficult to photograph. The SEM is the most
versatile tool available to the failure analyst to examine fracture surfaces at both low and high
magnification and to facilitate interpretations of fracture initiation sites (including inclusions),
fracture paths, and fracture morphologies. The ease of sample preparation, the wide range of
available magnifications, the large depth of field, and the versatility to capture photographs, all
combine to make the SEM the “instrument of choice” for most fractography.
3.1.4 Transmission Electron Microscopy
The operation of the transmission electron microscope (TEM) is very similar in principle to that
of the optical light microscope in transmitted light mode. Figure 3-15 compares the general
imaging path for the two instruments [3-14]. There are, of course, important practical
differences between the two microscopes. By definition, the sample in the TEM is viewed in
3-26
transmission mode, so that it must be sufficiently thin for the beam to penetrate completely. This
requirement for beam penetration of the sample complicates the applicability and ease of use of
the TEM beyond that of the SEM or OLM.
In transmission electron microscopy, a thin sample, typically less than 0.2 microns thick, is
bombarded by a highly focused beam of single-energy electrons. The beam has enough energy
for the electrons to be transmitted through the sample, and the transmitted or scattered electron
signal is greatly magnified by a series of electromagnetic lenses.
Due to the ease of scattering of electrons by air, the sample must be in a vacuum as is the case
for the SEM, although this is not a serious practical problem. The area that can be viewed is
small which restricts the TEM and places a unique burden on sample preparation. Since the eye
cannot see electrons, the image is generated by the action of the electron beam upon a
phosphorescent screen, on which the observations are made [3-13].
3.1.4.1 Magnification and Resolution Relative to OLM and SEM
As previously discussed, the principal advantage of the TEM and SEM compared to the OLM is
due to the shorter wavelength of the electron compared to visible light which extends the range
of magnification and resolution. Magnification possibilities on the order of 5,000,000X can be
inferred from fundamental principals for electron microscopes. In practice, aberrations in the
magnetic lenses limit the resolution of the TEM to 0.0002 to 0.0005 microns, at best. This is still
1000 times finer than that attainable with a light microscope and about 10 times better than the
SEM. This translates to a useful magnification of about 1,000,000X for the TEM. Such a high
magnification is difficult to obtain and more common magnifications are in the range of 10,000X
to 100,000X.
3-27
Figure 3-15
Comparison of the Imaging System for a Transmission Electron Microscope and a
Transmitted Light Optical Microscope [3-14]
3-28
Table 3-4
Comparison of Characteristics of Optical Microscopes, Transmission Electron
Microscopes, and Scanning Electron Microscopes [3-13]
3-29
Table 3-5
Comparison of Advantages and Disadvantages of Optical Microscopes, Transmission
Electron Microscopes, and Scanning Electron Microscopes [3-13]
Table 3-4 shows a comparison of the characteristics of optical microscopes, transmission

electron microscopes, and scanning electron microscopes while Table 3-5 shows relative
advantages and disadvantages of each [3-13].
3.1.4.2 Types of Samples and Sample Preparation
There are a number of drawbacks to the TEM technique. Many materials require extensive
sample preparation to produce a sample thin enough to be electron transparent, which makes
TEM analysis a relatively time consuming process with a low throughput of samples. The
structure of the sample may also be changed during the preparation process. Also the field of
view is relatively small, raising the possibility that the region analyzed may not be characteristic
of the whole sample.
Samples for the TEM are typically thin foils, replicas, or extraction replicas. A TEM specimen
must be approximately 0.1 micron or less in thickness in the area of interest. The entire specimen
must fit into a 3mm diameter cup. For thin foils, one approach uses a disc shaped sample with a
3-30
hole in the middle, the edges of the hole being thin enough for TEM viewing. The initial disk is
usually formed by cutting and grinding from bulk or thin film/substrate material, and the final
thinning done by ion milling.
The thin foil has the advantage that the sample is the actual material of interest which retains
relative spatial features whereas either replicas or extraction replicas involve transference of
features to a plastic film (e.g., cellulose acetate or polyvinyl alcohol) and subsequent steps before
viewing in the TEM. Therefore, artifacts are common in TEM samples, due both to the thinning
process and to changing the form of the original material. For example surface oxide films may
be introduced during ion milling and the strain state of a thin film may change if the substrate is
removed. Most artifacts can either be minimized by appropriate preparation techniques or be
systematically identified and separated from real information.
For fractography of microstructural features, a replica is usually used because it is essentially

impossible to prepare a thin film retaining the surface topology. In replication, a thin tape is
applied to the surface, softened to take on the detailed surface features, allowed to dry and
harden, and then the replica is stripped from the sample. This tape is too thick for viewing in the
TEM. Additionally, such polymeric materials are subject to damage from heating by the electron
beam. Instead it is common to take the tape replica and put a thin (0.02 microns) layer of carbon
on it in a vacuum coater. To enhance contrast of this carbon replica, a thin layer (0.02 microns)
of a heavy metal, such as chromium or platinum, is deposited, or shadowed onto the surface.
Then the acetate tape is dissolved in a suitable solvent, such as acetone, freeing the thin and
fragile replica into the solvent. This replica is removed onto a screen or grid, which is of suitable
size (about 3 mm diameter) to fit into the TEM. The replica is viewed through the openings in
the grid. The process of making such a replica is depicted in Figure 3-16.
3-31
Figure 3-16
Steps in Preparing an Indirect Carbon Replica using a Cellulose Acetate or Polyvinyl
Alcohol (PVA) Primary Replica [3-13]
It is possible to prepare a replica of the surface by depositing the carbon directly and not using
the tape. The carbon then is removed by adhesive tape, which is dissolved to free the replica.
This gives a replica of greatest fidelity having a resolution of about 0.002 microns. The two-
stage tape-carbon replica has a resolution of about 0.01 microns [3-13].
If it is desired to analyze particles on the surface, after the carbon layer is deposited, the
underlying metal (original specimen) can be dissolved chemically, freeing the replica with the
particles embedded in it. This is an extraction replica, often called a carbon extraction replica to
identify the role of the carbon deposition. An extraction replica allows analysis of the particles
in the TEM using electron diffraction and x-ray fluorescence. However, this process does
destroy the original metal surface. The preparation of an extraction replica is illustrated in Figure
3-17.
3-32
Figure 3-17
Schematic illustration of the Preparation of an Extraction Replica Used to Extract Particles
for Analysis in the TEM [3-13]
3.1.4.3 Viewing Modes
The TEM has a multiplicity of viewing and analysis modes with the two basic viewing modes
being brightfield and darkfield. A brightfield image is one using only the undiffracted (often
termed transmitted) beam to form the image. This is the counterpart to the brightfield image
discussed for the OLM. The size of the objective aperture in brightfield mode directly
determines the information to be emphasized in the final image. When the size is chosen so as to
exclude the diffracted beams, one has the configuration normally used for low-resolution defect
studies, so-called diffraction contrast. A darkfield image uses the diffracted beams to form the
image. Each beam can be selected individually by the objective aperture to form a darkfield
image which illuminates those parts of the image diffracting into this reflection.
In addition to brightfield and darkfield, the transmission electron microscope is capable of

performing a variety of other useful functions, thus expanding the flexibility with which
specimens may be critically examined. Some of the more common uses of the TEM include
3-33
electron diffraction, high resolution microscopy, tilting and stereo microscopy, low temperature
microscopy, energy loss spectroscopy, and x-ray microanalysis.
Complete to this Point as of February 20, 2007
3.2 Chemical Analysis
3.2.1 Bulk (Macro)
3.2.1.1 Conventional Quantitative Analysis – “Wet Chemical Analysis”
3.2.1.2 Spectrographic
3.2.1.2.1 Optical Emission
3.2.1.2.2 X-Ray Techniques (Energy and Wave Length Approaches)
3.2.1.3 Combustion and Other Methods (Carbon, Sulfur, Nitrogen, Hydrogen, Oxygen)
3.2.2 Local and Residual Level (Micro)
3.2.2.1 Spectrographic Using SEM or Microprobe (Energy Dispersive and Wavelength

Including Detectability and Accuracy)
3.2.2.2 Sample Dissolution and “Plasma” Detection Techniques (Flame, Inductively

Coupled Plasma)
3.2.3 Particle (Compound) Identification
3.2.3.1 X-Ray Diffraction
3.2.3.2 Electron Diffraction (TEM)
3.3 Mechanical Testing
3.3.1 Hardness
3.3.1.1 Macro
3-34
3.3.1.2 Micro
3.3.2 Tensile
3.3.3 Fracture
3.3.3.1 Charpy V-Notch
3.3.3.2 Plane Strain Fracture Toughness (KIc)
3.3.3.3 Elastic-Plastic Fracture Toughness (JIc)
3.3.3.4 Other Measure of Fracture Resistance
3.3.4 Creep-Rupture
3.3.5 Fatigue
3.3.6 Small Sample Testing Options
3.4 References
3-1 L. E. Samuels, Optical Microscopy of Carbon Steels, American Society for Metals,
3-2 Mortimer Abramowitz, Microscope: Basics and Beyond, Volume 1, Olympus America
Inc., 2003. Downloaded from http://olympusmicro.com.
3-3. Microscope General Catalog, Olympus America Inc., 2006. Downloaded from
http://www.olympus.co.jp/en/support/dl/catalog_new.cfm?genre_cd=310
3-4. G. L. Kehl, Principals of Metallographic Laboratory Practice, Third Edition, McGraw

Hill Book Company, New York, NY, 1949.
3-5. H. E. Rosenberger, “Differential Interference Contrast Microscopy,” Interpretive

Techniques for Microstructural Analysis, Ed. James L. McCall and P. M. French, pp. 79-
104, Plenum Press, New York, NY, 1977.
3-6. Optical Microscope Primer, Florida State University Website,

http://micro.magnet.fsu.edu, 2007.
3-35
3-7. W. J. Blaedel and V. M. Meloche, Elementary Quantitative Analysis – Theory and

Practice, Second Edition, Harper and Row, New York, NY, 1963.
3-8. Warren J. Smith, Modern Optical Engineering, McGraw-Hill Book Company, New
York, 1966.
3-9. Mike Meier, The Optical Microscope, University of California, Davis, Internet website
http://www.mat.sci.ucdavis.edu, 2004.
3-10. Donald S. Clark and Wilbur R. Varney, Physical Metallurgy for Engineers, D. Van
Nostrand Company, inc., New York, 1952.
3-11. ASME Specification SA-20, Specification for General Requirements for Steel Plates for
Pressure Vessels, Section II, Part A, Ferrous Materials Specifications. American Society
of Mechanical Engineers, New York, NY 2003.
3-12. Douglas C. Giancoli, Physics - Principles and Applications, Fifth Edition, Prentice Hall,
New Jersey, 1998.
3-13. Charlie R. Brooks and Ashok Choudhury, Metallurgical Failure Analysis, McGraw-Hill,
Inc., New York, 1993.
3-14. Metals Handbook, Eighth Edition – Volume 9: Fractography and Atlas of Fractographs,
American Society for Metals, Metals Park, OH: 1973.
3-36
4
SPECIAL TECHNIQUES
To Be Added in 2007
4.1 Ferrite Measurement

4.2 Phase and Particle Identification
Pearlite
Ferrite
Bainite
Martensite
Austenite
Sigma Phase
Carbides
X-Ray Diffraction
Electron Diffraction (TEM)
Nonmetallic Inclusions
4.3 Grain Size Measurements
4.4 Creep Void Assessment
Area Density
Grain Boundary Density
4.5 Oxygen Measurement in Weld Metals
4.6 Sensitization
4.7 Carbon Migration
4.8 Graphitization and Spheroidization
4.9 Carburization of Boiler Tubes During Service
4.10 X-Ray Diffraction
4-1
5
TYPICAL MATERIALS OF CONSTRUCTION
(Ferritics, Austenitics, Nickel-Based, Advanced Alloys)
To Be Added in 2007
5.1 Boiler
5.2 Piping
Fittings
Valves
5.3 Turbine
Casing
Rotors and Disks
5.4 Blading
5.5 Bolting
5-1
6
DAMAGE MECHANISMS
To Be Added in 2007
6.1 Tensile Overload

6.2 Creep
6.3 Continuum Damage Phase
6.4 Crack Growth Phase
6.5 Fatigue
6.6 Continuum Damage Phase
6.7 Crack Growth Phase
6.8 Creep-Fatigue
6.9 Environmental or Thermally Assisted
General Corrosion
Corrosion-Fatigue
Stress Corrosion Cracking
Sensitization
Sigma Phase
Graphitization and Spheroidization
Carburization
Carbon Migration
6-1
7
CHARACTERISTIC DAMAGE FOR SPECIFIC
COMPONENTS
To Be Added in 2007
7.1 Bolting
Operation at Temperatures where Creep and Stress Relaxation are Insignificant
Operation at High temperature when Creep and Stress Relaxation are Important
7.2 High Temperature Piping (Ferritic, Austenitic, and Advanced Materials)
Wrought Fittings and Associated Base Metal Concerns (Tees, Wyes, Elbows)
Girth Welds
Main Steam
Hot Reheat
Longitudinal Seam Welds
Main Steam
Hot Reheat
Attachment Welds
7.3 Tubing (Ferritic, Austenitic, and Advanced Materials)
Low Temperature
High Temperature
7.4 Low Temperature Headers and Steam Drums
Longitudinal and Girth Welds
Ligaments and Boreholes
Nozzles and Stub Tubes
Rolled
Welded
7.5 High Temperature Headers
Longitudinal Seam Welds
Girth Welds
Ligaments and Boreholes
Stub Tubes
7.6 HP/IP Turbines
Casings and Valve Bodies
Rotors
Disks
Blades (Rotating and Stationary)
7.7 LP Turbines
7-1
Characteristic Damage For Specific Components
Casings and Valve Bodies
Rotors
Disks
Integral
Shrunk-On
Blades (Rotating and Stationary)
7.8 Electrical Generators
Rotors
Retaining Rings
Conductors, Insulators, Slots, and Wedges
Coolers
7.9 Feedwater Heaters
Steam-Side Structures
Tubing
7.10 Condensers
7.11 Deaerators
7-2
Characteristic Damage For Specific Components
7-3
8
SUMMARY, CONCLUSIONS, RECOMMENDATIONS
To Be Added in 2007
8-1
A
COMMONLY USED ETCHANTS AND
PROCEDURES
A.1 Macroetches
A.1.1 Ferritic Steels
A-1
Commonly Used Etchants and Procedures
Table A- 1
Macroetches for Carbon Steel and Low Alloy Ferritic Steel
References
Metals Etchant Name Composition Uses/Remarks Source Original

25 mL HNO3 75 Use cold. Good for large sections that cannot be
Fe, Most Steels mL H2O immersed. Use with ground surfaces (Sauver). A-2 A-1
10-15 mL HNO3 85-
90 mL H2O or Use cold. Good general-purpose etch. Use with
Fe, Most Steels alcohol ground surfaces. A-2 A-1
Use at 160-180°F. Immerse sample for 10-60 min.
Developed by 1 Part HCl General-purpose etch. Use with cut or ground
Fe, Most Steels Uddeholm 1 part H2O surfaces. Most commonly used deep etch. A-2 A-1
38 mL HCl Use at 160-180°F. Immerse sample for 10-60 min.
Watertown 12 mL H2SO4 Good contrast obtained. Can be used for stainless
Fe, Most Steels Arsenal etch 50 mL H2O steel. Use with cut or ground surfaces. A-2 A-1
Solution A
5 mL HNO3
95 mL alcohol Etch 5 min. in Solution A, then dip 1 sec. in Solution
Solution B BN. Use polished sample. Good for revealing
10 mL HCl decarburization, carburization, and depth of
Fe, Most Steels 90 mL H20 hardening A-2 A-1
Saturated aq. picric
acid soln. plus Use polished surface. Etch up to 25 min. at room
Fe, Low-Alloy small amount of temperature. A few percent HCl can be added to
Steels wetting agent etch higher-alloy steels. Good for welds. A-2 A-1
Pour water into beaker, add the concentrated
30 mL H20 hydrochloric acid, then add the hydrogen peroxide.
20 mL HCl Use mixture within 20 minutes after preparing to
Carbon or Low- 10 mL of 30% assure proper strength. Good for locationing girth or
Alloy Steels 3-2-1 hydrogen peroxide long seam welds in piping sysytems.
Average ground or polished surface. Reveals
Carbon or Low- Nital, AWS 1-5% nitric acid carburization and decarburization, hardness
Alloy Steels D1.1 & D1.5 95-99% methanol penetration, cracks, segregation, weld features. A-3
CAUTION – These methods may be hazardous to personnel. They should not be

carried out by the uninitiated without professional supervision.
A-2
Table A-1 (Continued)
References
Carbon or Low- Picral, AWS 4 grams oicric acid Average ground or polished surface. Reveals
Alloy Steels D1.1 & D1.5 100 mL methanol carbon segregation, weld features. A-3
Heat solution to boiling (244-248°F), immerse
16 g. chromic acid speciment for 5-20 min., remove quickly and place
145 mL water in a container of cold water, rinse and dry. Detects
Carbon or Low- Alkaline 80 g. sodium oxygen rich regions which appear light against a
Alloy Steels Chromate hydroxide dark background. A-4 A-5
Immerse at room temperature, remove black
10 g. ammonium surface deposit by swabbing under water, wash and
Carbon or Low- Ammonium persulfate dry. An effective macroetrchant that can be used at
Alloy Steels Persulfate 100 mL water room temperature. A-4
100 mL sat. soln.
picric acid in water
several drops
sodium alkyl
sulfonate or Immerse at room temperature. Reveals the prior
Aqueous Picral 1 g. sodium austenite grain size in Q&T steels in some A-6,
Carbon or Low- (Bechet tridecylbenze circumstances. Also effective in revealing alloy A-7, A-
Alloy Steels Reagent) sulfonate segregation in Q&T steels A-4 8
1 mL bromine Immerse at room temperature. Identifies locations
Carbon or Low- Bromine- 100 mL methyl of solidification grain boundaries of austenite in
Alloy Steels Methyl Alcohol alcohol ferritic weld metal A-4
1g. cupric chloride,
50 g. ferric chloride, Immerse at room temperature for about 20 sec.
0.5 g. stannous Macroetching to reveal chemical segregation. It was
chloride, 30 mL. once thought that segregation of phosphorus
HCl, 500 mL. specifically was revealed, but it is now apparent that
Carbon or Low- Oberhoffers water, 500 mL. segregation of any of the common alloying
Alloy Steels Etch ethyl alcahol elements may be A-4 A-9
36 g. cupric Immerse specimen at room temperature for a
Carbon or Low- chloride, 145 mL. maximum time of 20 sec., rinse and dry. Reveals A-10, A-
Alloy Steels Fry's Reagent HCl, 80 mL. water localized regions of plastic deformation. A-4 11
Carbon or Low- Hydrochloric 100 mL. HCl heat solution 167-185°F. Etch by immersion for
Alloy Steels Acid (50%) 100 mL. water approx. 20 min. General macroetching. A-4
Dilute 1:1 with water immediately before use.
Immerse at room temperature and swab off the
carbonaceous deposit ocassionally. A macroetchant
Carbon or Low- Nitric-Acetic 50 mL. nitric acid effective for delineating the heat-affected zones
Alloy Steels Acid 50 mL. acetic acid adjacent to a weld. A-4
CAUTION – These methods may be hazardous to personnel. They should not be

carried out by the uninitiated without professional supervision.
A-3
A.1.2 Stainless Steels

Table A- 2
Macroetches for Stainless Steels and Similar High Alloy Materials
References

50 mL. HCl
25 ml. saturated
aqueous CuSO4
solution or 10g.
Austenitic Marbles CuSO4 in 50 mL.
Stainless Steels Reagent water Solution can be heated. Good general etchant. A-2 A-1
Attributed to Use electrolytically with polished surface. Platinum
Austenitic Miller and 10 g. CrO3 or stainless steel cathode, 60-100°F, 6 Volts dc, 1
Stainless Steels Houston 90 mL. water amp/sq. in., 2-7 min. Good for welds. A-2 A-1
Use electrolytically with polished surface. Platinum

or stainless steel cathode, 60-100°F, 6 Volts dc, 1
20 mL, ethylene amp/sq. in. Instantaneous contact made with
glycol specimen. Current density cut in half. Light
20 mL. HNO3 repolishing sometimes required before a 15-30 sec,.
Stainless Steel 100 mL. ethanol etch wit A-2 A-1
7.5 mL. water, 3.0

mL. HNO3, 1.0 mL. (a) Dissolve ammonium sulfate in water. (b)
HCl, 1.5 g. Dissolve ferric chloride in HCl. Mix (a) and (b); add
Nickel-Base ammonium sulfate, HNO3. Macroetch for Ni-base alloys. Make up fresh
Alloys Lepito's Etch 2.5 g. ferric chloride and swab on. Dilute with alcohol for microetch. A-12
30 mL. HCl
10 mL. HNO3 Warm to prevent pitting. Swab onto polished
Stainless Steel 10 g. FeCl3 surface. A-12
50 mL. H2O sat,
with CuCl2
50 mL. HCl
Stainless Steel 10 mL. HNO3 Swab onto polished surface. A-12
CAUTION – These methods may be hazardous to personnel. They should not be carried
out by the uninitiated without professional supervision.
A-4
A.2 Microetchs
A.2.1 Ferritic Steels

Table A- 3
General and Selective Microetchs for Carbon Steel and Low Alloy Ferritic Steel
References

1, 2, 3, or 5 mL. of nitric acid usual amount. The 2%
solution is most common; 5-10% used for high
alloy. Use by immersion of sample 5 to 60 s.
Darkens pearlite and gives contrast between
pearlite colonies. Etches the grain boundaries of
ferrite and attacks ferrite grains except those with
{100} parallel to the section surface. Reveals
Fe, C and low General 1-10 mL. nitric acid cementite mainly by developing relief between
alloy steels, cast Etchant-- 90-99 ml methanol cementite and the surrounding ferrite. Etches A-2,
iron Nital or ethanol untempered martensite well. A-4 A-1
Etch by immersion at room temperature for 5-60 s

or longer. Rinse and dry. Reveals cementite well in
a ferrite matrix. Does not develop grain boundaries
in ferrite, except very faintly after deep etching.
General Generally not suited for etching untempered
Fe, C and low Etchant -- 4 g picric acid martensite. Addition of about 0.5-1% zephiran A-2,
alloy steels Picral 100 mL ethanol chloride improves etch rate and uniformity (Igevski). A-4 A-1
General
Etchant -- Immerse at room temperature. Good for ferrite-
Vilella's carbide structures. Delineates the needles of plate
Reagent 100 mL. ethanol martensite. Useful for revealing the austenite grain
Low alloy steels (Picral- 5 mL. HCl size of Q&T steels in some circumstances. Outlines
(Q and Q&T) Hydrochloric) 1 g picric acid constituients in stainless steels A-2 A-1
An etchant developed for use in the investigation of

the structure of iron-carbon-chromium alloys. Used
to highlight carbides; in some alloys carbides may
Broad range of Selective 10 g K3Fe(CN)6 be stained or colored. Use fresh, hot or cold, up to
steels with Etchant -- 10 g KOH (or 7 g 10 min. When used cold, darkens chromium
carbide Murakani's NaOH), 100 mL. carbides and tungstides; cementite unattacked or
precipitates Reagent water barely attacked. When used hot, attacks cementite. A-2 A-1
Broad range of Selective
steels with Etchant --
carbide Blickwede and 1 g CrO3 Electrolytic etch, 2-3 V dc, 30 s. Carbides MC and
precipitates coworkers 100 mL. water M7C3 darkened; Mo2C outlined. A-2 A-1
Broad range of 10 mL. H2O2
steels with (30%)
carbide Selective 20 mL. 10% aq. Etch for 10 s at room temperature. Fe2MoC and
precipitates Etchant -- NaOH M6C outlined; M6C also colored. A-2 A-1
Broad range of Selective
steels with Etchant -- 4 g KMnO Outlines and colors carbides. Fe2MoC (blue) and
carbide Groesbeck's 4 g NaOH M6C (brown). Mo2C colored brown. (Fe,Cr)23C6
precipitates Reagent 100 mL. water attacked but (Fe,Mo)23C6 not attacked. A-2 A-1
A-5
A.2.2 Stainless Steels

Table A- 4
General and Selective Microetchs for Stainless Steels and Similar High Alloy Materials
References

Use at 6 V dc, 25 mm spacing. Carbides revealed
by etching for 15-30 s, grain boundaries after 45-60
s, sigma outlined after 6 s. 1-3 V also used.
Dissolves carbides, sigma strongly attacked,
austenite moderately attacked, feIrite unattacked.
To detect embrittlement [885°F (475°C)], use 5-7 V
Austenitic General 10 g oxalic acid dc, 0.1-0.2 A/cm2, 5-10 s. Unembrittled ferrite faint
Stainless Steels Etchant -- 100 mL. water bluish green, embrittled ferrite, brownish. A-2 A-1
General
Etchant -- 1 g picric acid Use at 20°C (68°F) for up to 1 min. Outlines
Austenitic Vilella's 5 mL. HCI constituents such as carbides, sigma phase, and
Stainless Steels Reagent l00 mL. ethanol delta ferrite; etches martensite. A-2 A-1
Solution A
25 mL. HCl
75 mL. ethanol For ferritic grades. Use polarized light and sensitive
Solution B tint. Sigma colored lemon-yellow to bluish white with
Austenitic Selective 10 mL. HCI stage rotation, ferrite and austenite remain maroon.
Stainless Steels Etchant -- 90 mL. water For austenitic grades, etch up to 30 min. A-2 A-1
Selective Use fresh. Difficult to dissolve at 20°C (68°F). Etch
Etchant -- at 20°C (68°F) for 3-5 s. Reveals sigma; carbide
Modified 30 g KOH and austenite unaffected. Use at 80°C (176°F) to
Austenitic Murakami 30 g K3Fe(CN)6, 60 boiling for 10-60 s. Sigma light blue, ferrite yellow,
Stainless Steels Reagent mL. water carbides barely attacked. A-2 A-1
Selective
Etchant --
Modified 30 g KOH Use at 90-95°C (194-203°F) for about 15 s.
Austenitic Murakami 30 g K3Fe(CN)6 Austenite light, ferrite tan, sigma reddish brown.
Stainless Steels Reagent 150 mL. water Can also be used at 20°C (68°F). A-2 A-1
1 part sat. aq.
KMnO4 For austenitic grades. Stains sigma pink, carbides
Austenitic Selective 1 part 8% aq. light brown. With deeper etch, sigma brown-orange,
Stainless Steels Etchant -- NaOH carbide green (uneven). A-2 A-1
Austenite colored first, then sigma, then carbide.
Selective Heat in air, 932- After 20 min, austenite mottled blue-green, sigma
Austenitic Etchant -- Heat l292°F (500-700° orange, carbide white. Very effective. Requires a
Stainless Steels Tint C) for up to 20 min. good polish. Light pre-etch sharpens resolution. A-2 A-1
Electrolytic etch at 1.5-3 V dc tor 3 s. Reveals
sigma (red-brown) and ferrite (bluish). Very
Selective effective. For austenitic PH grades, use 2V dc for 5
Etchant -- 10N KOH (56 g s; ferrite and sigma dark brown, α' brown to pale
Austenitic Potassium KOH in 100 mL blue to outlined. Ni3(Al,Ti) and Ni(AI,Ti) outlined.
Stainless Steels Hydroxide water) M23C6 pale yellow, austenite unattacked. A-2 A-1
A-6
A.3 References
A-1. G. F. Vander Voort, Metallography Principles and Practice, McGraw-Hill, New

York, NY, 1984
A-2. Boiler Tube Failure Metallurgical Guide, Volume2: Appendices. EPRI, Palo Alto,
CA: 1993. TR-102433-V2.
A-3. “Guideline for Etching Low-Alloy Steel,” American Welding Society Website,
http://www.aws.org/cgi-bin/mwf/topic_show.pl?pid=33548;hl=macroetch
A-4. L. E. Samuels, Optical Microscopy of Carbon Steels, American Society for

Metals, Metals Park, OH, 1980.
A-5. L. Fine, Metal Progress, 49 (1) p. 108, 1946.
A-6. S. Bechet and L. Beaujard, Rev. Met., 52, p. 830, 1955.
A-7. G. A. Dryer, D. E. Austin, and W. D. Smith, Metal Progress, 85 (1), p. 116, 1964.
A-8. D. R. Barraclough, Metallography, 6, p. 465, 1973.
A-9. J. S. Kiekaldy, J. von Destinon-Forstmann, and R. J. Brigham, Canadian Met.

Quart., 1, p. 59, 1962.
A-10. R. L. Bish, Metallography, 11, p. 215; 12, p. 147, 1979.
A-11. R. L. Bish, J. Aust. Inst. Metals, 21, p. 167, 1976.
A-12. Carl D. Lundin, Class Notes for Welding Metallurgy 5220, University of
Tennessee-Knoxville, 1984.
A-7
Export Control Restrictions The Electric Power Research Institute (EPRI)
Access to and use of EPRI Intellectual Property is The Electric Power Research Institute (EPRI), with
granted with the specific understanding and major locations in Palo Alto, California; Charlotte,
requirement that responsibility for ensuring full North Carolina; and Knoxville, Tennessee, was
compliance with all applicable U.S. and foreign export established in 1973 as an independent, nonprofit
laws and regulations is being undertaken by you and center for public interest energy and environmental
your company. This includes an obligation to ensure research. EPRI brings together members, participants,
that any individual receiving access hereunder who is the Institute's scientists and engineers, and other
not a U.S. citizen or permanent U.S. resident is leading experts to work collaboratively on solutions to
permitted access under applicable U.S. and foreign the challenges of electric power. These solutions span
export laws and regulations. In the event you are nearly every area of electricity generation, delivery,
uncertain whether you or your company may lawfully and use, including health, safety, and environment.
obtain access to this EPRI Intellectual Property, you EPRI's members represent over 90% of the electricity
acknowledge that it is your obligation to consult with generated in the United States. International
your company’s legal counsel to determine whether participation represents nearly 15% of EPRI's total
this access is lawful. Although EPRI may make research, development, and demonstration program.
available on a case-by-case basis an informal
assessment of the applicable U.S. export classification Together…Shaping the Future of Electricity
for specific EPRI Intellectual Property, you and your
company acknowledge that this assessment is solely
for informational purposes and not for reliance
purposes. You and your company acknowledge that it
is still the obligation of you and your company to make
your own assessment of the applicable U.S. export
classification and ensure compliance accordingly. You
and your company understand and acknowledge your
obligations to make a prompt report to EPRI and the
appropriate authorities regarding any access to or use
of EPRI Intellectual Property hereunder that may be in
violation of applicable U.S. or foreign export laws or
regulations.
© 2007 Electric Power Research Institute (EPRI), Inc. All rights reserved.
Electric Power Research Institute, EPRI, and TOGETHER…SHAPING
THE FUTURE OF ELECTRICITY are registered service marks of the
Electric Power Research Institute, Inc.
Printed on recycled paper in the United States of America 1012744
Electric Power Research Institute

3420 Hillview Avenue, Palo Alto, California 94304-1338 • PO Box 10412, Palo Alto, California 94303-0813 • USA
800.313.3774 • 650.855.2121 • askepri@epri.com • www.epri.com

EPRI - Fossil Power Plant Components Failure Analysis Guideline

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EPRI - Fossil Power Plant Components Failure Analysis Guideline

Uploaded by

Copyright:

Available Formats

Fossil Power Plant Components

Failure Analysis Guideline

Technical Update, March 2007

EPRI Project Manager

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I)

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT

B. W. Roberts Engineering Consulting

This report was prepared by

B. W. Roberts Engineering Consulting

Results and Findings

Challenges and Objectives

Applications, Values, and Use

2 SEQUENTIAL STEPS TYPICALLY INVOLVED IN A FAILURE ANALYSIS........................2-1

3 TYPICAL LABORATORY METHODS AND EQUIPMENT UTILIZED ...................................3-1

5 TYPICAL MATERIALS OF CONSTRUCTION.......................................................................5-1

6 DAMAGE MECHANISMS ......................................................................................................6-1

7 CHARACTERISTIC DAMAGE FOR SPECIFIC COMPONENTS ..........................................7-1

8 SUMMARY, CONCLUSIONS, RECOMMENDATIONS .........................................................8-1

A COMMONLY USED ETCHANTS AND PROCEDURES ...................................................... A-1

The eleven steps typically involved in a failure analysis are as follows:

2.1 Collection of Failure Background Data

Design Operations Engineering

Fabrication Maintenance Management

FABRICATION & ERECTION

2.1.1 Literature Review Relevant to the Failed Component

Information that is useful to a failure investigation is often organized as “component specific”

2.1.2 Literature Review Relevant to the Postulated Failure Mechanism(s)

Failure mechanisms are typically discussed in metallurgical and mechanical engineering

Whether organized as component specific or by failure mechanism, similar options to those

2.2 Photographic Documentation and Identification

2.3 Nondestructive Examination

Nondestructive examination (NDE) is very important in the life-cycle management of power

Liquid Penetrant Testing

Magnetic Particle Testing

Testing Agency Qualification

Commonly used nondestructive methods of examination of failed components include:

Magnetic Particle Testing

Liquid Penetrant Testing

trapped in discontinuities, and as a contrasting background to enhance the visibility of penetrant

Eddy Current Testing

ET is most widely applied to non-ferromagnetic materials, although continuing advances have

Portable Alloy Analysis

Material confirmation is an essential element in any failure analysis. Complete confirmation

There are four approaches to the measurement of ferrite:

2.4 Preliminary Examination of Failed Component

The significance of any indications of abnormal conditions or abuse in service should be

2.5 Sample Removal, Identification, and Preservation of Essential Features

Sample identification, either beforehand or immediately following sample removal, is essential.

2.5.1 Metallographic Samples

2.5.2 Chemical Analysis Samples

2.5.3 Mechanical Testing Samples

2.6 Metallographic Examination

Macroscopic examinations generally cover the magnification from 1X (visual examination) to

2.6.1.1 Visual, Low Power Microscopic, SEM

The amount of information that can be obtained from low-magnification examination of a

2.6.1.3 Polishing and Etching

Even in the absence of a specific metallurgical imperfection, examination of metallographic

2.6.2.1 Polishing and Etching

Polished metal specimens usually show no structural characteristics. However, on occasion

degree of conformance of the failed component to the requirements of the specification. It is

2.7 Chemical Analysis

In any failure investigation, it is strongly recommended that routine chemical analysis be