THESIS

IIIIIII

      

 

 

IIIIIIIIIIIIIIIIIIIII I III IIIIIIIIIIII III

293 01565 6956

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

Examination of Dislocation Structures in
B2 Stiochiometric NiAl Alloys using TEM and SEM Techniques

presented by

'Boon-Chai Ng

has been accepted towards fulfillment
of the requirements for

 

Master's degree in Materials Science

%%

Major professor

 

 

Date May 6, 1997

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

 

 

 

PLACE N RETURN BOX to remove We checkout trom your record.
TO AVOID F INES return on or before date due.

    
   

DATE DUE DATE DUE DATE DUE

I I-[Q
IIQJI II :I
:I IQ
QI‘IIfi.

MSU le An Affirmative ActlonlEquel Opportunity lnetlttllon

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EXAMINATION OF DISLOCATION STRUCTURES IN
BZ STOICI-IIOMETRIC NiAl ALLOYS USING TEM AND SEM TECHNIQUES

By

Boon-Chai Ng

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTERS OF SCIENCE

Department of Materials Science and Mechanics

1997

ABSTRACT

EXAMINATION OF DISLOCATION STRUCTURES IN
B2 STOICHIOMETRIC NiAl ALLOYS USING TEM AND SEM TECHNIQUES

By

Boon-Chai Ng

Toughened and embrittled single crystals of stoichiometric NiAl alloys were
examined using TEM and SEM. It is observed that air cooling has a profound impact on
the yield strength and the mobility of the dislocations whereas furnace cooling results in
severe embrittlement. Furnace cooled specimens exhibit dislocations that are heavily
tangled whereas air cooled specimens exhibit much fewer dislocation tangles. The
electron channeling contrast imaging (ECCI), a scanning electron microscopy technique,
is used to examine the dislocation behavior near crack tips of 4-point bend specimens.
ECCI imaging of the dislocations ahead of the crack tips show that the number of
dislocations and their distribution ahead of the crack tips for air cooled specimens are
much more than that of the furnace cooled specimens. This is because there are more
mobile dislocations and fewer pinned dislocations in the air cooled specimens, hence
pre-existing and newly generated dislocations will be able to move farther away from the

crack tip than when they are pinned due to furnace cooling.

ACIG‘IOWLEDGEMENTS

This thesis would never have been made possible if not for the many individuals
who were there to help and guide me in one way or another. Among the many
individuals, I would like to single out a few to express my sincere thanks.

I would like to thank my advisor, Dr. Martin A. Crimp for his guidance and for his
many helpful suggestions and painstaking reviews during the preparation of this thesis.
His dedication, easy going and ever believing that I “will always do a good job” have
been an inspiration to me. A fine mentor!

I would like to thank the Office of Naval Research and Dr. George Yoder
(Scientific Officer) for their funding and support (Grant No. N00014-94-1-0204), the
National Science Foundation (Grant No. DMR#9302040) and Michigan State University
for providing the funds for the purchase of the CamScan 44 FE PEG-SEM used in this
study.

Thanks to Dr. Ram Darolia at General Electric Aircraft Engine Division for the
supply of NiAl single crystals.

Special mention goes to Benjamin Andrews Simkin, who was instrumental in my

acquiring the ECCI technique, and Biman Ghosh, for his assistance in analyzing the

iii

dislocation structures. Special thanks to Alan Gibson for proof reading the first draft of
this paper.
Last but not least, thanks to members of my thesis committee, Dr. D.S. Grummon

and Dr. T.R. Bielcr for their time and helpful suggestions.

TABLE OF CONTENT

 

 

 

 

 

 

 

LIST OF TABLES vi
LIST OF FIGURES vii
CHAPTER 1: INTRODUCTION 1
1.1 INTRODUCTION .................................................................................................................. 1
1.2 RECENT STUDIES ................................................................................................................ 6
1.3 SLIP SYSTEMS IN NIAL ..................................................................................................... 16
1.4 OBSERVATION OF DISLOCATIONS IN ENHANCED PLASTICITY IN NIAL .............................. 17
1.5 OBJECTIVE ....................................................................................................................... 18
CHAPTER 2: TEM EXAMINATION 19
2.1 EXPERIMENTAL PROCEDURE ........................................................................................... 19
2.2 SLIP TRACE ANALYSIS ..................................................................................................... 22
2.3 DISLOCATION ANALYSIS .................................................................................................. 25
2.4 RESULTS .......................................................................................................................... 25
2.4.1 UNDEFORMED SPECIMENS ...................................................................................... 28
2.4.2 DEFORMED SPECIMENS ........................................................................................... 32

2.5 COMPUTER SIMULATION .................................................................................................. 45
2.6 SUMMARY ....................................................................................................................... 51
CHAPTER 3: SEM EXAMINATION ' 53
3.1 INTRODUCTION TO ELECTRON CHANNELING CONTRAST IMAGING (ECCI) ...................... 53
3.2 EXPERIMENTAL PROCEDURE ........................................................................................... 56
3.3 RESULTS ......................................................................................................................... 60
3.4 SUMMARY ....................................................................................................................... 75
CHAPTER 4: DISCUSSION 76
CHAPTER 5: GENERAL CONCLUSION _ 80
REFERENCES 82

 

LIST OF TABLES

Table 1: Properties of test specimens ...................................................................................... 27
Table 2: Burgers vector analyzed for each of the tested specimens in

correlation to their heat treatment histories and yield strength ................................................ 38
Table 3: Parameters used in simulating dislocation “A” ......................................................... 50

LIST OF FIGURES

Figure 1: BZ structure is the ordered bcc lattice. Two kinds of atoms (A and

B) occupy comer positions or body center position respectively. ........................

Figure 2: Phase diagram illustrating the atomic structure as a fimction of

temperature for the Ni-Al alloy system ..............................................................

Figure 3: Plot of thermal conductivity for Ni-SOAI in comparison to a typical

Ni-based superalloy as a function of temperature. Adapted from [2]. ..................

Figure 4: Cracks propagate by a combination of plastic and elastic process.
Plastic portion of the crack is created by the dislocations. Elastic portion of

the crack is a result of brittle fracture of the dfz. ..................................................

Figure 5: The emitted dislocations exert a back stress to the crack tip thus
reducing the local stress intensity factor k. Stress relaxation is viewed as a
result of an increase of the total number of dislocations within the plastic zone.

Figure 6: Crack, the dislocation free zone and plastic zone in an infinite
elastic medium. f(x) is the distribution function of the dislocations. Adapted

from [22] ...............................................................................................................
Figure 7: Orientation of specimen within the sterographic triangle. ...................

Figure 8: Flow chart of the heat treatment procedure. ........................................

Figure 9: Two faces of a deformed specimen showing the slip planes with

measured angles OI and B .......................................................................................

Figure 10: Schematic representation of a compression sample displaying the

measurement of OI and [3 used in determining the Slip plane. ...............................

Figure 11: Sterographic projection of the specimen displaying the angles a

and [3 used to determine the slip plane. .................................................................

Figure 12: Burgers vector analysis for undeforrned 673FC specimen. The

same area is shown with different operative reflections[011], [200] and [110]

vii

..................... 3

..................... 4

..................... 5

................... 12

................... 14

................... 18
................... 20

................... 21

................... 23

................... 24

................... 26

................... 29

Figure 13: Microsgraphs of a) 473FC specimen and b) 673AC specimens
displaying both <001> and <011> dislocations ...................................................................... 30

Figure 14: Single dislocations that are widely spaced with few dislocation
loops and tangles were observed in a) undeforrned specimen, b) undeforrned
673FC specimen and c) undeforrned 673AC specimen ........................................................... 31

Figure 15: Burgers vector analysis for deformed 673AC specimen shows both
“a” dislocaitons with b= [ 1 O 0 ] and “b” dislocations with b= [ O l 1]
present. The same region with different Operative refelctions area shown:

a)[110],b)[200],c)[211] ....................................................................................... 33

Figure 16: Burgers vector analysis revealed both b==[ I O O ] and h = [ 0 1 I ]
dislocations present in deformed 673FC specimen. The_ same area is shown
with different operative reflections a) [ l l 0 ], b) [O l 1 ], c) [O 2 O] ................................ 34

Figure 17: Micrographs of deformed 473FC specimen displaying both <001>
and <011) dislocations. ‘a’ denotes <001> and ‘b’ denotes <011>. ....................................... 35

Figure 18: <001> and <011> dislocations were observed in deformed 1590FC
specimen. ‘a’ denotes <011> and ‘b’ denoted <011> .............................................................. 36

Figure 19: Burgers vector analysis reveals both <001> and <01 1>
dislocations present in deformed 1590AC specimens. ‘a’ denotes <001> and
‘b’ denotes <011>. ................................................................................................................... 37

Figure 20: Dislocations were observed to be heavily tangled in deformed
473FC specimen. a) brightfield image and b) weak beam image ........................................... 40

Figure 21: Dislocations were Observed to be heavily in deformed 1590FC
specimen. a) brightfield image and b) weak beam image ...................................................... 41

Figure 22: Dislocations were Observed to have less tangles in deformed
673F C specimen. a) brightfield image and b) weak beam image ........................................... 42

Figure 23: Very few dislocation tangles were observed in deformed4159OAC
specimens. a) brightfield image and b) weak beam image ..................................................... 43

Figure 24: Very few dislocation tangles were Obsreved in dformed 1590AC
specimens. a) brightfield image and b) weak beam image ..................................................... 44

Figure 25: Dislocation mark “A” in an undeforrned 673AC specimen which
will be used to illustrate the technique of defect identification by image
matching ................................................................................................................................... 48

Figure 26: A set of experimental images (a) - (g) of dislocation “a” of 673AC
speciemen take with the diffraction condiitons given in Table 3 and four sets

viii

of computed images for the same diffraction conditions corresponding to the

burgers 1/2 [101] and 1A: [110] .................................................................................................... 49
Figure 27: Near surface defects will cause a change in the bse yield due to

variation from the bragg condition .......................................................................................... 55
Figure 28: Selected area channeling pattern (SACP) .............................................................. 57

Figure 29: Ray diagrams illustrating schematically the scanning action in the
selected area channeling pattern (SACP) ................................................................................. 58

Figure 30: Schematic diagram of a 4-point bend sample showing the notch
plane and the notch direction ................................................................................................... 59

Figure 31: Schematic diagram showing the notched sample in the 4-point
bend fixture attached to the E. Fullan deformation stage. ...................................................... 61

Figure 32: SACP showing a) a zone axis, b) tilting to one of the channeling
band c) final adjustment of aligning the edge of the band (Bragg condition) to
the microscope axis .................................................................................................................. 62

Figure 33 Crack tip in deformed specimen 673AC under different imaging
conditions. a) high magnification SB image of the crack tip region. inserted
image shows a low magnification SE image of the secondary crack and
highlighting the area analyze. b) ECCI image of the crack tip region showing
dislocaitons lying end-on . ....................................................................................................... 64

Figure 34 Examples of dislocation distribution left in the wake of the cracks
for 1590FC specimen. a) SE image of the crack edge, b) ECCI image of the
same crack edge region. .......................................................................................................... 65

Figure 35: A higher bse yield resulted when the rastering beam is near to a
edge of a specimen. .................................................................................................................. 67

Figure 36: Schematic diagram showing how the combination of mode I and II
tear results in a change in the BSE yield for both sides of the crack edge ............................... 68

Figure 37: Images of a crack edge of a 673AC specimen illustrating a) SE
image and b) ECCI image with dislocations in dark/bright contrast lying end-
on .............................................................................................................................................. 69

Figure 38: Images of an area 80 um from the edge of the crack for specimen
673AC before (a) and after (b) 4 point bend test. no significant change in the
dislocation density was observed. ............................................................................................ 71

Figure 39: Comparsion of dislocation densities for a) 673AC specimens, b)
673FC specimen and c) 1590FC specimen ............................................................................ 72

ix

Figure 40: Changes in contrast of dislocation images when the channeling
band is changed from a) +g to b) - g ........................................................................................ 73

Figure 41: Cracks do no show any change in contrast as the channeling
condition changes from + g to - g ............................................................................................ 74

CHAPTER 1

INTRODUCTION

1.1 Introduction

One of the greatest challenges currently facing the materials community is the
need to develop a new generation of materials to replace Ni-based superalloys in the hot
section of gas-turbine engines for aircraft-propulsion systems. The present alloys, which
have a Ni-based solid solution matrix surrounding Ni3Al-based precipitates, are currently
used at temperatures exceeding 1373 K. Since Ni3Al melts at 1669 K and Ni at 1726 K,
it is clear that significantly higher Operating temperatures with improvement in efficiency
and thrust-to-weight ratio can only be attained by the development of an entirely new
material system [1]. This problem is a primary reason for the current high level of
interest in high temperature intermetallic compounds.

A high melting temperature is the first and most obvious requirement for the new
generation of materials since the melting temperature must exceed the operating
temperature. Low density is very important since the weight of the propulsion system as
a whole decreases rapidly with decreasing Iweight of the rotating components [2].
Oxidation resistance at high temperature is also critical as are cost considerations. In

addition, the material must possess some intrinsic plasticity and toughness (the required

levels of which are not yet totally clear [1]). However, some plasticity at room
temperature is required for fit-up of the components during assembly.

Intermetallic compounds have the requisite combination of properties to satisfy
some of the list Of requirements just described. An interrnetallic compound is a true
compound of two or more metals that has a distinctive structure in which the metallic
constituents are in relatively fixed stoichiometic ratios and are usually ordered on two or
more sublattices, each with its own distinct population of atoms [3]. Among the
interrnetallic compounds currently being studied, NiAl has been the subject of the most
study and development.

NiAl possesses the ordered cubic B2 (space group Pm 3m, CsCl prototype)
crystal structure. This structure consists of two interpenetrating primitive cubic cells
(Figure 1), where Al atoms occupy the cube comers of one sublattice and Ni atoms
occupy the cube comers of the second sublattice. The Ni-Al binary equilibrium phase
diagram [4] is shown in Figure 2. NiAl exhibits a wide single phase field, and the
stoichiometric composition melts congruently at 1911 K.

NiAl offers many advantages: (1) density of 5.95 g/cm3, approximately 2/3 that
of Ni-based superalloys, (2) thermal conductivity that is four to eight times that of Ni-
based superalloys (Figure 3), (3) excellent oxidation resistance, (4) simple, ordered
body-centered-cubic derivative B2, (CsCl) crystal structure, (5) lower ductile to brittle
transition temperature (DBTT) relative to other interrnetallic alloys, (6) high melting
temperature that is approximately 573 K higher that Ni-based superalloys, (7) relatively

easy process-ability by conventional melting, and powder-and metal forming techniques,

 

 

 

 

 

 

 

 

 

OA OB

Figure 1. B2 structure is the ordered bcc lattice. Two kinds of atoms (A and B) occupy
comer positions or body center position respectively.

Atomic Percent Nickel

 

 

 

 

 

 

 

 

 

 

 

 

 

u to to so u so on to no to Ice
.m‘ I ‘1 v 1 ' I" 1' 1' l I J 1' L
I
Ieoo-j L
I I
"w Aim
9 ' (m)
2 15°01 ..
3 ' ‘ m
8
E W
O 4
P 1
”01 mm
\\ ‘ 1:11:13
100- .- J.‘
I” 3 i’ MN an:
I (Al) ‘ ‘ 5
“on Iii 1'0 3'. lo 30 ee 7'0 e'e' 03 Ice
Al Weight Percent Nickel NI

Figure 2. Phase diagram illustrating the atomic structure as a function of temperature for
the Ni-Al alloy system. Adapted from [4].

 

80.0—

 

60.0“

40.0“1

N i-based Superalloy

20.0m

 

Thermal Conductivity (W/cm K)

I. l l ~l l l
I I I r I T

'200 400 600 800 1000 1200 1400

 

 

Figure 3. Plot of thermal conductivity for Ni-50Al in comparison to a typical Ni-based
superalloy as a function of temperature. Adapted from [2].

and (8) high Young's modulus relative to Ni-based superalloys. For a review of the
properties of NiAl, see ref. [5].

The use of NiAl as structural members, however, suffers from two major drawbacks:
(1) poor ductility and toughness at ambient temperatures and (2) low strength and creep at
elevated temperatures [6,7]. Because of the excellent high-temperature capability of NiAl,
considerable effort has been devoted to understanding brittle fracture and improving

mechanical properties of NiAl during the past years. These will be covered in the following

pages.

1.2 Recent Studies of NiAl

Several recent studies have shown that it is possible to improve the room temperature
properties of near-stoichiometric NiAl single crystals.

Hack et al. [8,9] have recently reported some interesting results that contribute to our
understanding of deformation and fracture in single crystal NiAl. Their studies indicate that
the fracture resistance of single crystal NiAl can be dramatically improved by controlled heat
treatments. NiAl single crystals that were furnace-cooled from a homogenization treatment
at 1590 K exhibited a tensile elongation of 1.0% and a fracture toughness of 2.4 MPa.m"2 at
room temperature. The room temperature ductility increased to 7.0% and the toughness
increased to 16.7 MPam”2 when these crystals were
reheated to 673 K followed by air cooling. The beneficial effect of the 673 K heat treatment
disappeared as the specimens were cooled slowly inside a furnace. The authors attribute the
low ductility and poor toughness of single crystal NiAl to strain-aging embrittlement

involving the pinning of mobile dislocations by interstitials such as carbon and oxygen,

similar to the strain-aging embrittlement in mild steels [10]. The strong interaction between
the interstitial C atoms and the cores of dislocations results in a potent strengthening effect
for C. At elevated temperatures, the free energy of the solution of C in Fe is lowered by the
random distribution of C in solution. As the temperature is lowered, the driving force for the
C atoms to diffuse to dislocation cores increases so as to lower the internal energy. If the
temperature is lowered rapidly enough to limit redistribution of the C atoms by diffusion, the
C remains randomly distributed at low temperatures. However, if the cooling rate allows for
ample interstitial diffusion of the C, or the temperature is maintained at a level where the
driving force for redistribution is high and diffusion is sufficiently rapid, the bulk of the C
atoms can assume positions at or near the cores of dislocations, forming so-called Cottrell
atmospheres [11]. The interstitial content in these NiAl crystals is reported to be only about
100 wt. ppm carbon and 50 wt. ppm oxygen. The data [8,9] further suggests that ductility
and toughness of NiAl are more strongly dependent upon mobile dislocation density rather
than on the inherent mobility of dislocations in the ordered lattice.

The population of mobile dislocations can be affect by the strong interaction between
the interstitial solute atoms and the highly strained region at the core of a dislocation [12].
Decoration of the core region with interstitial solutes renders the dislocations immobile, thus
reducing the mobile dislocation density. This phenomena is known as strain aging. The
strain aging embrittlement in NiAl has been studied by Brzeski et al. [13]. In their studies,
compressive load-strain curves for test specimens showed smooth yielding at room
temperature and serrated yielding at temperatures between 100 and 200°C during
compression testing with the serrations beginning with the onset of the plastic deformation.

This is typical of solute drag interactions between the dislocations and interstitial solutes [12]

and is known as dynamic strain aging. It is characterized by the repetitive breaking away
from the atmosphere by a dislocation followed by diffusion of the atmosphere to the
dislocation where it pins the dislocations again. This leads to the observed serrated yielding
in load-strain curves [13].

Strain aging embrittlement is also consistent with observations of recovery of the
flow stress and a sharp yield drop in polycrystal NiAl which has been pre-deformed by
hydrostatic pressure. Margevicius and Lewandowski [14,15] suppressed premature fracture
in tension and compression tests of NiAl alloys by the use of superimposed hydrostatic
pressure. Both the tension and compression tests, conducted at 0.1 MPa following
pressurization at 500 MPa., reveal a significant effect of pressurization on the subsequent
flow stress at 0.1 MPa. The tests showed that flow stresses obtained following pressurization
are significantly lower than those obtained on specimens simply tested at 0.1MPa. Such a
decrease in flow stress is consistent with the pressure-induced generation of mobile
dislocations, either at second phase particles or grain boundaries. Margevicius and
Lewandowski [14,15] concluded that dislocations were generated at second phase particles
(e.g. impurities or inclusions) in the NiAl or alternatively, at grain boundaries due to the
elastic anisotropy of MA]. In addition to pressure lowering the initial flow stress, specimens
tested in tension with 500 MPa pressure exhibited fracture strains in excess of 10% compared
to the usual 1-2% [8]. While the pressure-induced generation of mobile dislocations
probably explains the former, superimposed pressure greatly reduces the rate of damage

accumulation [16], thereby increasing the ductility.

Based on the embrittling mechanism proposed by Hack et a1. [8,9], Liu [17], reported
that the fracture resistance of NiAl crystals is expected to be effectively improved by either
reducing the interstitial content or scavenging interstitial by certain alloying additions.

Microalloying with Fe, Ga, or M0 at levels less than 0.5 atomic % has been reported
to increase the room temperature ductility of <110> crystals by as much as 6% [18].
Microalloying with Fe also improves the ductility of <111> crystals and lowers the ductility
to brittle transition temperature (DBTT) for <001> crystals. The mechanism by which Fe,
Mo and Ga additions enhance the room temperature ductility is not known, although it may
be an indirect effect of point defects or impurity gettering by the small additions differences .

While microalloying, pre-straining and alternative heat treatments clearly cause
alterations in the initial state of NiAl crystals, there are no changes in operative slip systems
and/or significant changes in dislocation substructure following deformation. So it is unclear
whether the improvements in ductility and toughness arise from changes in the intrinsic
mobility of the dislocations, the mobile dislocation density, and/or from dislocation
generation and distribution at crack tip interactions which delay the brittle fracture process.

More recently, Levit et a1. [19] showed that it is possible to achieve a high tensile
elongation of ~25% by optimizing factors such as impurity content, heat treatment,
orientation, prestraining and surface condition of NiAl single crystals. In their experiments,
homogenized high purity, stoichiometric single crystals were prestrained (compressed) at
1273 K before undergoing tensile test under ambient conditions. These high temperature
prestraining leads to an increase in the tensile elongation from an average of 16% to 26%.

The authors concluded that the high temperature prestraining resulted in the production of a

IO

dislocation-free subgrain volumes, separated by low energy dislocation boundaries. These
subgrain volumes act as a source for mobile dislocations.

These recent studies have lead to the proposition that NiAl and BCC metals having
levels of interstitial impurities, are subject to a strain aging phenomenon caused by elastic
interactions between the interstitial atoms and dislocations.

The term strain aging indicates a time-dependent strengthening or hardening process
resulting from elastic interaction of solute atoms with the strain fields of dislocations in
plastically deformed metals and alloys [20]. Aging reactions can occur in either static or
dynamic modes depending upon whether they occur prior to or during plastic deformation.
As discussed by Weaver [21], static strain aging typically occurs in metals and alloys
following prestraining, unloading (either partially or fully), aging for a prescribed time and
then reloading at the same strain rate as the prestrain. It is typically manifested by an
increase in yield stress or flow stress following aging and the return of a sharp yield point in
the deformed alloy. Dynamic strain aging, the result of interactions between diffusing solute
atoms and mobile dislocations during plastic deformation, is manifested by the appearance Of
serrations, load drops, jerkiness or other discontinuities in the stress-strain curves obtained in
constant-extension-rate tensile or compression tests. These phenomena are associated with
the dynamic formation and migration of solute atmospheres around dislocations during
deformation.

The brittle fiacture of structural materials have also been studied using in-situ
straining TEM. Many authors [22-24] have used in—situ straining TEM to observe
dislocations near the crack tips in the hope of characterizing the role of dislocations at the

crack tip regions and how the crack tip stress field can be modified by these dislocations.

The limitation of these studies is the constraints imposed by the requisite thin foils. This
present study uses electron channeling contrast imaging (ECCI) to image dislocations at the
crack tips of bulk specimens. This technique has gain attention as a potential method for
imaging near surface crystal defects of bulk specimens and will be describe in detail in
chapter 3. In order to explain the significance of the ECCI studies, a a brief review of crack
tip deformation and its impact on the dislocation theory of fracture is useful.

In general, dislocations are emitted from the crack tips during the early stages of crack
propagation and are driven out of the crack tip area, leaving behind a dislocation-free zone
(DFZ) (Figure 4). In metals with low stacking fault energies the dislocations pile up in the
form of linear arrays on a slip plane which is coplanar with the crack. In metals with high
stacking fault energies, the dislocations cross slip out of the original slip planes and form a
broad plastic zone. The cracks propagate by a combination of plastic and elastic processes in
which the plastic portion of the crack opening is created by the dislocations that were emitted
from the crack tip while the elastic process occurs as a result of brittle fracture of the
dislocation-free zone (DFZ) along the slip planes. Afier emitting a number of dislocations,
the crack propagates along the slip planes of these dislocations. As the crack propagates, the

crack tip continues to emit dislocations and these dislocations move ahead of the crack.

12

crack tip "L

]_ i
\ _L dislocations
i I .L
dislocation _J_ 'J—
free zone (dfz) _I_

I

 

 

Figure 4. Cracks propagate by a combination of plastic and elastic process. Plastic
portion of the crack is created by the dislocations. Elastic portion of the crack is a result
of brittle fracture of the dfz

13

When a dislocation is emitted from a crack tip, it exerts a back stress to the crack tip
thus reducing the local stress intensity factor k (Figure 5). This phenomenon is known as the
dislocation shielding of the crack tip [25,26]. Figure 6 shows a schematic diagram of a crack,
the dislocation free zone and plastic zone in an infinite elastic medium. As mentioned by
Chang and 0hr [22], if the applied stress is fixed, the increase of the total number of
dislocations, which is essentially equal to the crack Opening displacement (COD),
corresponds to a decrease of the length of the DFZ (e/c) and an increase of the length of the
plastic zone (a/c).

where COD= (4co/1ru) ln(a/c),

and o = friction stress in the plastic zone

u = shear modulus

a = length of the plastic zone
2c= length of crack

e = length of DFZ

Thus the stress relaxation can be viewed as a result of an increase in the total number
of dislocations within the plastic zone. The local k is very sensitive to the length of the DFZ
[22,23]. When the DFZ is absent, the shielding of the crack tip by the dislocations is
complete and hence the local k is reduced to zero. In this situation, no thermodynamic
driving force exists for this crack to propagate. For a complete review of the crack tip

deformation and its impact on the dislocation theory fracture, see ref. 22-24.

l4

 

b k
crack tip ac stress—L _|_
3 _I_
Applie _I_
force W "L "L dislocations
‘— _I_
. . i 1
local stress 1nten51ty _L
factor k

 

Figure 5. The emitted dislocations exert a back stress to the crack tip thus
reducing the local stress intensity factor k. Stress relaxation is viewed as a result of an
increase of the total number of dislocations within the plastic zone.

15

 

 

DFZ

 

Figure 6. Crack, the dislocation free zone and plastic zone in an infinite elastic medium.
f(x) is the distribution function of the dislocations. Adapted from [22].

l6

Slip Systems in MA]

Critical to understanding the mechanical behavior Of any material is the study of the
dislocations and slip systems of the material.

In general, <100> dislocations are the most commonly observed dislocations in
deformed NiAl single crystals [19,27-30]. These <100> dislocations have been Observed to
slip on both {011} and {001} planes. The critical resolved shear stress (CRSS) calculation
based on 0.2% yield stress, for slip on the <100> {011} and <100>{001} systems, are similar
[31] and results in the active slip plane being a function of single crystal orientation.

While <100> dislocations are the most commonly Observed dislocations in deformed
NiAl, <111> dislocations have also received considerable attention. This is due to the
expectation that activation of the <111> slip might result in enhanced ductility in polycrystal
NiAl as <111> slip is capable of satisfying the Von Mises criteria of five independent slip
systems. Based on this premise, many current research efforts have been concentrated on
encouraging slip of the <111> type. These <111> dislocations have been activated in single
crystals by some researchers [32,33] by inhibiting the slip of <100> dislocations. Slip of
<100> dislocations may be inhibited by orienting to the <001> “hard” orientation where the
resolved shear stress in the <100> slip direction is very low and only at or below room
temperature. <111> slip has also been achieved by alloying with Cr and Mn [34,35].
However it should be noted that the generation of <111> dislocations has never resulted in
increased plasticity [34].

<110> dislocations have been observed for deformation temperatures above 873 K
[32,35,36]. Like <111> dislocations at lower temperature, these <110> dislocations can only

be produced by inhibiting the slip of <100> dislocations by orienting single crystals tO <001>

17

hard orientations. At present there is still some questions as to what extent these <110>
dislocations contribute to plastic strain. Fraser et a1. [37] have argued that these <110>
dislocations, the products of the interaction of two <100> dislocations, are sessile and do not
contribute to the plasticity of NiAl. Field et al.[31] on the other hand showed that <110>
dislocations are actually responsible for plasticity and that <100> dislocations are debris
resulting from a slip dissociation of<110> screws at superjog and that it is primary the debris

that is being Observed, and not the actual slipping dislocations.

1.4 Observations of dislocations in enhanced plasticity NiAl

As outlined in the section on dislocations above, there has been some success in
enhancing the ambient temperature ductility and toughness of NiAl. These studies have used
microalloying, controlled thermal treatment and prestressing to presumably increase
dislocation mobility and/or generation. However, these studies have been complimented
with only limited observations of the dislocation structures associated with enhanced
plasticity.

Standard diffraction contrast (g.b) experiments carried out by Levit et a1. [19]
revealed that essentially all the dislocations had <100> Burgers vectors.

In the case of improved tensile elongation as a result of microalloying, weak beam
TEM analysis revealed no differences between alloyed materials and binary materials tested
in compression [18]. In both cases, similar structures, dominated by <100> dislocations were
Observed. The dislocations produced as a result of hydrostatic prestress were also reported to
be <100> type dislocations [14,15]. However, the analysis presented in the study is not

entirely clear.

18

With regards to the strain age phenomenon reported by Hack’s group [8,9], analysis

of the dislocation structures has not been performed.

1.5 Objectives

It is the objective of this study to examine the types of dislocations, their behavior, as
well as the dislocation distribution in the region of the crack tips and crack edges of both
toughened and embrittled single crystal NiAl. This has been carried out using the
transmission electron microscopy (TEM) as well as scanning electron microscopy (SEM).

Since no dislocation structures were analyzed in Hack et al.’s [8,9] studies, this study
examined the role of dislocations and how it affects static strain aging in the brittle behavior
of NiAl. The heat treatment processes as outlined by Hack et al.’s [8,9] for toughened and
embrittled conditions of single crystal NiAl were used in order to duplicate similar conditions
as reported in their studies. After heat treatment, selected specimens were subjected to room
temperature compression to 2.5% plastic strain to induce dislocation generation. These
deformed specimens have been prepared for examination in the TEM. The results have been
compared to undeforrned specimens. Other specimens have been examined employing the
electron channeling contrast imaging (ECCI) technique performed using SEM to Observe

dislocation behavior and generation in 4-point bend specimens in the region of crack tips.

CHAPTER 2

TEM EXAMINATION

2.1 Experimental Procedure

Commercial purity stoichiometric single crystal NiAl was oriented by x-
ray diffraction using the back-reflection Laue technique to lie within the stereographic
triangle as shown in Figure 7. Specimens measuring 4.5x4.5x10 mm were cut using a
high speed diamond blade wafering saw. In order to duplicate the embrittled and
toughened conditions as reported by Hack et al. [8,9], these specimens were then given a
homogenization anneal at 1590 K for 48 hours in a Centorr Vacuum Furnace followed by
either furnace cooling with a cooling rate of ~0.1 K/s (specimens denoted as 1590FC) or
air cooling, with a cooling rate of ~1 K/s. Samples that were air-cooled were reheated to
1573 K in the same furnace for 3 hours and air cooled (1590AC). Some of these 1590AC
were then reheated to 473 K for 90 minutes and furnace cooled (473FC). Samples
(1590FC) that were firrnace cooled afier the initial homogenized anneal were reheated to
673 K in a Lindberg furnace for 12 hours and were either air-cooled (673AC) or fumace
cooled (673FC). The embrittled specimens are 1590FC, 473FC and 673FC whereas the
toughened specimens are 673AC and 1590AC. The flow chart of the various heat

treatment processes is shown in Figure 8.

l9

20

 

001 ‘ 011

Figure 7. Orientation of specimen within the stereographic triangle.

21

 

48 hours in a vaccum atmosphere

L1590 K homogenized anneal for 1

 

ir l
a coo ed fumace cooled

   

 

 

 

reheat to 1590 K reheat to 673 K
for 3 hours for 12 hours

 

 

 

   

 

reheat to 473 K
for 1.5 hours

 

 

 

 

 

473 FC 1590AC 673 AC 673FC . 1590FC

Figure 8. Flow chart of the heat treatment procedure. Adapted from [8,9].

22

Some of these heat treated samples were then mechanically ground and polished
to 0.3 pm using A1203 polishing medium. These polished specimens were then placed in
a compression fixture and subjected to room temperature compression using an Instron
Machine model 4206. These samples were compressed at a nominal strain rate of 10“ per
second to 2.5% plastic strain to induce dislocation generation and slip. Using optical
microscopy, the slip lines were observed and the active slip planes were determined by
standard slip trace analysis which will be explained in section 2.2.

Undeformed and deformed samples were then cut into 3 mm diameter rods and
sectioned parallel to the slip planes using electro-spark machining. These discs were then
ground to 0.3 m using grit SiC paper and electropolished using a twin-jet polisher. The
electrolyte consisted 0 one part nitric acid (HNO3) to three parts methanol.
Electropolishing was performed at a temperature of 243 K and a voltage of 12 V.
Dislocation observation was performed using a H-800 Hitachi TEM. Burger’s vectors of

the dislocations were determined using the standard g0b=0 invisibility criteria.

2.2 Slip Trace Analysis

Following compression, the specimens were examined optically to observe
dislocation slip traces. An example of slip traces on two perpendicular faces is Shown in
Figure 9. The active slip planes were determined by trace analysis. An example of which
is given as follows. The angles a and B were measured on a pair of micrographs showing

the front and right faces of the rectangular compression sample as shown in Figure 10.

23

 

Figure 9. Two faces of a deformed specimen showing the slip planes with measured
angles a and B.

24

Compression axis

 

 

Figure 10. Schematic representation Of a compression sample displaying the
measurement of a and B used in determining the slip plane.

25

The angles were then plotted on the stereographic projection of the front face in the
manner shown in Figure 11. The two angles define the plane which causes the slip trace
on the single crystal. The plane and its pole (90 degree from the plane) are shown on the
stereographic projection. In all cases, regardless of heat treatment and cooling rates,

{100} was analyzed as the slip plane.

2.3 Dislocation analysis

Dislocation Observation and analysis was performed on a Hitachi H-800 200KV
transmission electron microscope using the brightfield as well as weak beam imaging.
To Obtain a sufficient number of g vectors for contrast analysis, the primary poles, {001 },
{011} and {111} were used. As NiAl is an anisotropic crystal, the dislocations may not
be completely invisible when g-b=0 and g- bxu=0 conditions are satisfied [38]. This is
because no planes remain undistorted around edge or screw dislocations because of the
anisotropy. Thus the invisibility method for determining b often relies on finding
diffraction vectors which result in weak contrast rather than complete invisibility.
Computer simulation was used where possible, to confirm the Burger’s vectors. This will

be described in section 2.5.

2.4 Results

Data obtained from the compression tests are shown in Table 1. These show the
yield strength of each specimen tested, along fracture toughness values from Hack et a1.
[8,9] corresponding to the same heat treatments. It is Observed that a slow cooling rate

(furnace cooled) tends to increase the yield strength but correspondingly results in lower

26

 

 

 

Figure 11. Sterographic projection of the specimen displaying the angles OI and B used to
determine the slip plane.

27

Table 1. Properties of test specimens

 

 

 

 

 

 

 

Specimen Condition of specimen Yield *Fracture
Stress Toughness
MPa MPa.m“2
473FC 1590 K homogenization anneal and air 400 2.8
cooled. Reheated to 1573 K and air cooled.
Then reheated to 473 K and furnace cooled.
1590FC 1590 K homogenization anneal and furnace 450 2.4/4.0
cooled.
673FC 1590 K homogenization anneal and fumace 360 5.8
cooled. Reheated to 673 K and furnace
cooled.
1590AC 1590 K homogenization anneal and air 260 15.6
cooled.
673AC 1590 K homogenization anneal and furnace 150 16.7

 

cooled. Reheated to 673 K and air cooled.

 

 

 

 

*Adapted from Hack et a1. [8,9]

 

28

fracture toughness, whereas fast cooling (air cooled) tends to decrease the yield strength
but results in a much higher fracture toughness value. The dislocation structures of each
specimen were examined and the Burgers vectors b were determined using the g - b=0
invisibility criterion. Generally dislocations will be invisible when their b lies in the
reflecting plane. Using two-beam conditions for a series of individual reflections, two
such conditions of invisibility should be found. As b is common to both reflections, b

must be the zone axis Of the two planes.

2.4.1 Undeformed Specimens

Burgers vector characterization of an undeforrned 673FC specimen is shown in
fig. 12. The “a” dislocations with b = [ 0 0 1 ] are observed to be in weak contrast for
both [2 0 0 ] and [ 1 1 O ] reflections while the “b” dislocations b= [ 0 I l ] are observed
to be in weak contrast in both [2 0 0 ] and [ 0 1 1 ] reflections.

Figure 13 shows micrographs of undeforrned 473FC and 673AC specimens
exhibiting <001> and <011> dislocations. Likewise, undeforrned 1590AC and 1590FC
specimens show similar types of dislocations. It is noted that majority of the dislocations
displayed in all the undeforrned specimens are dominantly <001> type with a few
scattered <011> dislocations.

The gross morphologies of all the undeformed specimens show similar dislocation
densities. Figure 14 shows micrographs of 473FC, 673FC and 673AC specimens

showing single dislocations widely spaced with few dislocation loops and tangles.

29

 

IL; \ I ‘.\ \ H

 

Figure 12. Burgers vector analysis for undeforrned 673FC specimen. The same area is
shown with different operative reflections: [011], [200] and [110]. ‘a’ denotes <001>
and ‘b’ denotes <011>.

30

 

Figure 13. Micrographs of a) 473FC specimen and b) 673AC specimen displaying both
<001> and <011> dislocations. ‘a’ denotes <001> and ‘b’ denotes <011>.

31

 

 

 

Figure 14. Single dislocations that are widely spaced with few dislocation loops and
tangles were observed in a) undeforrned 473FC specimen, b) undeformed 673F C
specimen and c) undeformed 673AC specimen.

32

2.4.2 Deformed Specimens

Figure 15 shows a Burgers vector analysis for a deformed 673AC specimen,
which posses a low yield stress value of lSOMPa. The “b” dislocations b = [ 0 l I ] are
Observed to be in complete extinction in the [ 2 0 0 ] reflection and weak in [ 2 I I ]
reflection. The “a” dislocations are analyzed as b = [ 1 0 0 ] which displays
extinction/weak contrast in both [ 0 2 0 ] and [0 I 1 ].

The deformed 673FC specimen, which possess a higher yield stress (300MPa.)
than the 673AC specimen, is shown in Figure 16. Burgers vector analysis show that
dislocations with b = [ I 0 0 ] are observed to be in extinction for [ 0 1 I ] reflection and
in weak contrast for [ 0 2 O ] reflection whereas dislocations with b = [ 0 1 I ] are
observed to be in weak contrast for both [ 2 0 0 ] and [ 2 l 1 ] reflections.

Deformed 473FC, 1590FC, and 1590AC specimens were also analyzed as having
both < 0 0 1 > and < 0 1 1 > dislocations (see Figures 17-19 respectively).

Based on these Observations, regardless of heat treatment, cooling rates, and
corresponding yield strength, it has been determined that the dislocations exhibit similar
Burgers vectors, i.e. < 0 0 1 > and < O 1 1 >, although primarily <001> dislocations with
scattered <011> dislocations. Dislocations were observed to move under the influence of
the electron beam indicating that although these materials do not display extensive
ductility, the dislocations in the structures are mobile. Table 2 summarized Burgers
vectors analyzed for each of the tested specimens in relation to their heat treatment
histories and yield strength.

The gross morphology of each Of the specimens was examined to determine the

type of dislocation structures present. Using both bright field and weak beam imaging it

33

 

 

Figure 15. Burgers vector analysis for deformed 673AC specimen shows both “a”
dislocations with b= [ 1 0 0 ] and “b” dislocations with b=[ 0 1 1 ] present.

The s_arn_e region with different operative reflections are shown:

a) [ 1 10], b) [200], c)[211].

34

 

Figure 16. Burgers vector analysis revealed both b= [ i o o ] and b=[ o 1 i ]
dislocations present in deformed 673FC specimen. The same area is shown with difl‘erent
operativereflections a) [ 110], b) [01 1], c) [020].

35

. “a. b “9
fl .
b/
1' / y
5'
.9
' " 1 pm

Figure 17. Micrographs of deformed 473FC specimen displaying both <001> and <011>
dislocations. ‘a’ denotes <001> and ‘b’ denotes <011>.

36

 

Figure 18. <001> and <011> dislocations were observed in deformed 1590FC specimen.
‘a’ denotes <001> and ‘b’ denotes <011>.

37

 

Figure 19. Burgers vector analysis reveals both <001> and <011> dislocations present in
deformed 1590AC specimen. ‘a’ denotes <001> and ‘b’ denotes <011>.

38

Table 2. Burgers vectors analyzed for each of the tested specimens in correlation to
their heat treatment histories and yield strength.

 

 

 

 

 

 

 

Specimens Yield Stress Burgers vector b
MPa
473FC 400 Primarily <001> Scattered <011>
1590FC 450 Primarily <001> Scattered <011>
673FC 360 Primarily <001> Scattered <011>
1590AC 200 Primarily <001> Scattered <011>
673AC 150 Primarily <001> Scattered <011>

 

 

 

 

 

39

was observed that deformed 473F C and 1590FC specimens, having the highest yield
strength compared to the rest of the samples, show dislocations that are heavily tangled
(Figures 20 and 21 respectively). Dislocations, which moved under the influence of the
electron beam, were observed to be restrained when they moved into these tangles. These
dislocation tangles thus appear to act as obstacles to dislocation motion. The 673FC
specimen, displaying fewer dislocation tangles as compared to both 473F C and 1590F C
specimens, is shown in Figure 22. 1590AC specimen (Figure 23) and 673AC specimen
(Figure 24), which were toughened and displayed lower yield strengths, were observed to
have very few dislocation tangles and as such, dislocations were able to move without
many obstructions.

From this examination, it is concluded that deformed specimens having higher
yield stresses (and hence lower fracture toughness) tend to have dislocations that are
heavily tangled, whereas specimens having lower yield strengths (and hence higher
toughness) tend to have fewer dislocation tangles. These dislocation tangles are a result
of deformation following the compression tests as undeforrned specimens do not display
dislocations that are heavily tangled. Deformed specimens displaying heavily tangled
dislocations had undergone furnace cooling. On the other hand, deformed specimens
which do not Show much dislocation tangles had been rapidly cooled. This leads to the
possibility that with slow cooling, there is sufficient time for interstitial to positions at or
near the cores of dislocations and results in strain aging embrittlement [13]. Such

behavior is consistent with static strain aging reported by Weaver [21] who concluded

40

 

Figure 20. Dislocations were observed to be heavily tangled in deformed 473FC
specimen. a) brightfield image and b) weak beam image.

41

 

Figure 21. Dislocations were observed to be heavily tangled in deformed 1590FC
specimen. a) brightfield image and b) weak beam image.

42

 

Figure 22. Dislocations were observed to have less tangles in deformed 673FC specimen.
a) brightfield image and b) weak beam image.

43

 

Figure 23. Very few dislocation tangles were observed in deformed 1590AC specimen.
a) brightfield image and b) weak beam image.

 

Figure 24. Very few dislocation tangles were observed in deformed 673AC specimen.
a) brightfield image and b) weak beam image.

45

that it is the interstitial C that is responsible for yield stress obtained for furnace cooled
specimens. Fast cooling on the other hand would ‘trap’ the interstitials in the solution.
As a result the pre-existing dislocations are not pinned and so more mobile dislocations
are available. This would result in the lower yield stress obtained for air-cooled
specimens. High yield stress (or low fracture toughness or a lack of ductility) then must
have derive from the low mobility of dislocations.

The observation of heavily tangled dislocations in deformed specimens that are
furnace cooled (slow cooling) have led to the conclusion that static strain aging occurs as
a result of having sufficient time for the interstitial to diffuse into the cores of dislocaitons
and pinning them. Hence these observations support Hack et al.’s [8,9] conclusion that
the brittle behavior in single crystal NiAl were a result of a static strain aging induced

reduction in mobile dislocation density.

2.5 Computer Simulation

The use of experimentally invisible or weak images to detect diffracting
conditions in which g-b=0 is open to the criticism that dislocations in elastically
anisotropic materials Often display residual contrast. On the other hand, the invisibility
criteria is simple to use and there seems to be no doubt that in the majority of cases
examined, experimental invisibility coincide with g- b=0. As most materials are usually
anisotropic and dislocations are in general neither of screw nor edge orientation, true
invisibility can be difficult to achieve experimentally. For this reason, an image matching
technique is sometimes used to confirm Burgers vectors. This technique is based on

computer simulation of a series of dislocation images and matching these simulated

46

images with experimental images. The computer program used in the present study,
which was adapted from Head et a1. [39], is based on the dynamical theory of image
contrast.

The general method of identification of defects using the computed micrographs
may be stated as follows; experimental micrographs of the unknown defects are taken
under different diffracting conditions, e. g. different diffracting vectors taken in different
electron beam directions. Then informed guesses are made as to what the unknown
defect might be and theoretical micrographs are computed for these under the same
diffracting conditions as the experimental micrographs. When the theoretical and
experimental micrographs match consistently, the defect is identified.

Unlike the use of invisibility criteria, this technique of computerized image
matching does not involve only the choice of weak or invisible images. On the contrary,
the images which are well suited to the comparison technique are often those which have
easily distinguished topological features; that is, images with ‘character’.

In order to Obtain the data necessary for accurate simulation, micrographs and
their corresponding diffracting patterns should be collected in such a way that the
crystallography of the defect and of the foil, as well as the diffracting conditions for each
micrograph, can be specified as completely as possible. The procedure for obtaining this
experimental information involves taking, for the same defect, a series of electron
micrographs and corresponding selected area diffraction patterns for a number of
different beam directions and diffracting conditions. It is important to recognize the
kikuchi line patterns associated with low index crystallographic directions for the

material under consideration. Important data includes the operative diffracting vector g

 

47

for each micrograph, the beam direction B, the dislocation line direction u, the foil
normal F, a value of the deviation from the Bragg condition w, the foil thickness t, and
the anomalous absorption coefficient. Methods of obtaining the above parameters are
given in ref. [39].

The first step in determining a Burgers vector by image computation involves
choosing a set of experimental images of the dislocation from the available micrographs.
Figure 25 shows a dislocation marked ‘A’ in an undeforrned 673AC specimen. This
dislocation image will be used to illustrate the technique of defect identification by image
matching. It should be noted that the matching progress will be greatly facilitated if the
images have distinctly differing topologies arising from distinguishing features of
contrast. Moreover, the chosen set should contain three non-coplanar diffracting vectors
in order to sample all components of the displacement field of the defect.

A set of micrographs of dislocation ‘A’ are shown in the first column of Figure 26
and labeled (a) - (g). The diffracting vectors g, beam directions B and values of the
deviation from the Bragg condition w, corresponding to each micrograph are listed in
Table 3 together with the appropriate values of theoretical two-beam extinction distances
and the values of apparent anomalous absorption coefficient. In addition to the set of
seven experimental images, four corresponding sets of computed images for different
possible Burgers vectors are presented. These images have been computed for the
diffracting conditions specified by the values given in Table 3. The object is to select the
set of computed images that is in best visual agreement with the experimental set. As
shown, the computed images for Burgers vector [ 1 0 0 ] and '/2 [ 1 1 0 ] are not in

agreement with the experimental image (c) and the computed images with Burgers vector

48

 

Figure 25. Dislocation marked ‘A’ in an undeforrned 673AC specimen which will be
used to illustrate the technique of defect identification by image matching.

49

.952? E80? Bomam 05 8 womenoamaoo 802280 dogma can 08 new women: @8388 mo 38 Sam was
m 033. 5 53m Eowmvaoo gauge 05 "EB :83 .<. nausea—mu. .«o g - A8 women: Engage mo 8m < .3 053m
A: VN: A8V~> AooV ASov 09:5 3885qu

... ....

I xii. III - I I we?

  

'1' (I.-. , /l./I.’! a v---"

Q

L I
e -I ,I. . \wt‘ --.I .
I 7... _,I 1“
LIL -- :I I

I
I

I
I

50

Table 3. Parameters used in simulating dislocation ‘A’

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

gvector 200 110 020 '200 011 3 30 031
LU 3236 3236 3236 3236 3236 3236 3236
LBM 016 3215 3019 0911 333 3515 499
LFN 3112 3112 3112 3112 3112 3112 3112
w 0.8 0.85 0.6 0.4 0.7 0.6 0.26
t 1.3 3.5 4 1.3 3 3 2.16

where

LU = direction of dislocation line

LBM = electron beam direction

LFN = foil normal

w = deviation from Bragg condition

I

Other parameters used include:

Elastic constants: C11 = 2.03

= foil thickness in extinction distances

C12 = 1.34

Anomalous absorption coefficient : ANO = 0.07

C44= 1.16 [48]

 

51

[ 0 0 l ] is not in agreement with the experimental image ((1). On these ground, the
Burgers vector of dislocation ‘A’ is unlikely to be either of these possibilities and a
further comparison of the remaining computed and experimental images confirms this
conclusion. However, it can be seen that visual agreement is sufficient to conclude that

the most likely Burgers vector for dislocation ‘A’ is V2 [101].

2.6 Summary

Mechanical testing (compression) of heat treated single crystal NiAl alloys shows
that the yield stress increases as the cooling rate through the critical temperature region
decreases.

The slip trace analysis shows ( 0 0 1 ) to be the glide plane for all the specimens.

TEM examination of both deformed as well as undeforrned specimens show the
following observations and analysis;

1. Both < 0 0 1 > and < 0 1 1 > types of dislocations were observed although
dominantly <0 0 1> type.

2. Irrespective of heat treatment histories and yield strengths, dislocations were
observed to move under the influence of the electron beam suggesting that the
dislocations in the structure are mobile.

3. Undeformed specimens, irrespective of heat treatment histories, display similar
dislocation densities. Single dislocations were observed to be widely spaced
with few dislocation tangles.

4. Deformed specimens having higher yield stress (and hence lower fracture

toughness) display extensive dislocation tangles whereas those with lower

52

yield stress (and hence higher fracture toughness) show fairly straight
dislocations. These dislocation tangles are a result of deformation during
compression test.

. In the deformed specimens, dislocations were observed to be restrained when
they moved into dislocations that were tangled. These dislocation tangles act
as an obstacle to dislocation motion and are a hindrance to the movement of
the mobile dislocations in the structure.

. Deformed specimens displaying heavily tangled dislocations had undergone
furnace cooling whereas deformed specimens which do not show much
dislocation tangles had been rapidly cooled. This leads to the possibility that
with slow cooling, there is sufficient time for interstitial atoms to diffuse to
positions at or near the cores of dislocations and results in strain aging
embrittlement. As a result, the pre-existing dislocations are immobilized. This

influenced the mobile dislocation density.

CHAPTER 3

SEM EXAMINATION

3.1 Introduction to electron channeling contrast imaging (ECCI)

This section deals with examining the brittleness problem in NiAl. This problem
may be associated with difficulty in the generation of fresh dislocations and the ability to
move these dislocations, particularly at the highly stressed regions of crack tips. To
understand this dislocation generation and mobility, the present study examined the
dislocation behavior near crack tips using electron channeling contrast imaging (ECCI), a
scanning electron microscope (SEM) technique. This technique has gained attention as a
potential method for imaging near surface crystal defects (for reviews, see ref. 40-44).
Although in—situ straining transmission electron microscopy (TEM) [45-47] and atomic
force microscopy (AF M) [48] have in the past been used to study the dislocation behavior
near crack tips, the ECCI technique provides certain characteristic features that are
unmatched by both TEM and AF M. ECCI allows experiments to be performed on bulk
specimens as opposed to TEM. Thus ECCI has the potential to overcome the difficulties
associated with the thin foils used in TEM, including the complications of specimen
preparation, surface effects coupled with the strong interaction between the crack tips and

dislocations, and the unknown stress state at the crack tip. The limitation of the AF M

53

54

approach to assessing dislocations near crack tip is that it estimates the number of
dislocations within a slip plane by measuring the slip step height as a function of distance
from the crack tip [48] and is thus an indirect approach to studying dislocations. It does
not provide the kind of visual information about the dislocations near the crack tips that is
revealed by the ECCI technique.

The ECCI technique may be best understood by examining the source of the
contrast, namely electron channeling. The basis behind electron channeling is that the
backseattered electron (BSE) yield from a perfect crystal is dependent on the angle of
incidence of the electron beam relative to the numerous Bragg conditions within the
crystal. The BSE yield undergoes a significant change upon crossing the Bragg condition
[49-51]. When the general bulk specimen is tilted such that it is at one of the Bragg
conditions, then any near surface defects will cause a change in the BSE yield, due to
variations from the Bragg conditions (Figure 27). The BSE yield from the tilted volume
will be different from the rest of the crystal, thus highlighting the local strain field (for
instance) of a defect. The reason for the near surface sensitivity is that the electrons from
the incident beam loose coherency in penetrating the crystal, and provide less contrast the
farther in they. travel, limiting the depth of detection [52,53]. The basis for ECCI is thus
somewhat analogous to the imaging of dislocations in scanning transmission electron
microscopy (STEM), or the diffracting plane tilt model of dislocation contrast for the
TEM, in that the contrast from defects will give rise to changes in the local diffracting

conditions.

55

   
 
  

BSE Yield

 

 

 

 

E E Tilt volume

93 is the Bragg angle

Incident
electron
beam

Specimen

a is slightly greater or less than 93

Figure 27. Near surface defects will cause a change in the BSE yield due to variations

from the Bragg condition.

56

Electron channeling patterns (ECPs) are formed by rocking the electron beam
about two perpendicular axes on the sample and displaying the resultant BSE signal as a
function of tilt angle (Figure 28). This can be accomplished by using a dedicated set of
scan coils to rock the beam through the desired range of angles while it is confined to a
very small area on the specimen (Figure 29). A pattern captured in this way is referred to
as a Selected Area Channeling Pattern (SACP). Actual ECCI imaging is performed using
a rastering beam but SACPs are necessary to set up the channeling conditions for ECCI,
much like a selected area diffraction pattern is used to set up a two beam diffracting

condition for TEM imaging.

3.2 Experimental Procedure

Commercial purity single crystals of NiAl were oriented to {0 l 0} using both
Laue back-reflection x-ray and SACP, and then cut to nominal dimensions of 4 x 4 x 20
m using electro-discharge machining, EDM. The specimens were then subjected to a
homogenization heat treatment at 1590 K for 48 hours followed by a slow furnace cool
(specimens denoted as 1590FC). Some of these specimens were then retreated at 673 K
for 12 hours and either fumace cooled (673F C) or air cooled (673AC). The specimens
were mechanically ground and polished to 0.3 pm using SiC and A1203 polishing media.
The samples were then notched in a { 0 1 0 } < 1 0 0 > fashion using EDM, where the
given plane represents the plane of the notch and the given direction denotes the direction

of the notch (Figure 30).

57

 

Eo<mv Sosa mfifioqqfio moan vogue—om .wm oBmE

 

 

83.5.," 2955

 

    

e 0.95 was—cox

3 295 $5.03.

.58 3885:)— m

 

58

 
  

SACP aperture

\

lens

—
upper scan coils I

 

lower scan coils

lens

 

specimen

 

Figure 29. Ray diagrams illustrating schematically the scanning action in the selected
area channeling pattern (SACP).

59

notch plane

 

notch direction

=>

20mm

4mm

 

4mm

Figure 30. Schematic diagram of a 4 point bend sample showing the notch plane and the
notch direction.

60

The notched specimens were electropolished at 195 K in a solution of 90v.%
methanol and 10v.% nitric acid for approximately 60 minutes. Each specimen was then
placed in a specially designed 4 point bend fixture attached to a E. Fullam deformation
stage (Figure 31) mounted in a CamScan 44FE F EG-SEM. Bending was performed at
the minimum attainable strain rate.

ECCI was performed on each sample prior to loading to assess the initial
dislocation densities and distributions. Following bending strain, the samples were
reassessed to determine the nature of the dislocation distribution. Particular attention was
paid to areas near crack tips and along the edges of cracks.

In order to set up channeling conditions for ECCI imaging, the electron beam was
rocked about a point on the specimen surface to obtain a SACP. Using the obtained
SACP, the specimen was tilted to one of the channeling bands. Final adjustment was
made by placing the edge of the band (which is at the Bragg condition [43,54]) onto the
microscopic axis (see Figure 32) before switching to the imaging mode. In the present
study, ECCI was conducted employing { 1 l 0 } and { 2 0 0 } channeling bands for

CODII'flSt.

3.3 Results
Specimens deformed using in-situ 4 point bend testing did not exhibit any visible
crack extension prior to fracture. In all cases, no crack initiation or propagation was

observed prior to catastrophic failure, which occurred via rapid crack motion. As such,

61

.03 senescence 833m .m a 3 3583
ounce EB “58-... 05 8:0 332 0388 05 wgogm 583% oumaonom 4m 059m

 

HOHOEOmGOHXO

 

Esta
28 334

 

 

    

 

\\\\\\\\\.\\\\ ‘\\\\\\\\\\
\\\\\\\\\\\\\\\\\\\\\\\\

\\\\\\\\\\\\\\\\\\\\\\\\

\\\\\\\\\\\\\\\\\\\\\\\\
‘

 

 

 

EflSfiooE How

038% @380: =8 32

 

 

 

62

.38 unconfined: 05 9 308950 $85 33 2:
mo owvo 0%. ma??? mo E253? Ham 6 £33 @5585 05 mo 28 8 man—u 3 £me 88 Am wagon,“ mo<m .3 onE

AT.

Mic mqouiu}: .

\

 

63

observations of crack tip regions were made on secondary cracks protruding out from the
main fracture crack.

A crack tip in a deformed 673AC specimen is shown in Figure 33 under different
imaging conditions. The inserted image shown in Figure 33a is a low magnification
secondary electron image (SEI) showing the secondary crack and highlighting the region
to be analyzed. Due to the lack of topographic features, the SE image of the crack tip in
Figure 33a shows low contrast, whereas the ECCI image of the same crack region in
Figure 33b shows dislocations lying roughly end-on with one end of each dislocation
exiting the crystal. These dislocations appear somewhat similar to the contrast displayed
by end-on dislocations in thin TEM foils [55]. The dislocations decrease in contrast as
they penetrate deeper into the sample. This is because the electrons from the incident
beam loose coherency in penetrating the crystal and hence provide less contrast the
farther in they travel. In this sample, dislocations are observed only within a few microns
of the crack tip and edges. These dislocations serve to shield the crack tip and hence
increase the macroscopic fracture toughness of the material. The crack in the ECCI
image appears different from the SE image because electrons will continue to penetrate
deeper into the crack with little or no backscattered electrons escaping the crystal leading
to a low BSE signal.

Figure 34 shows SE and ECCI images of a crack edge in a 1590FC specimen.
The ECCI image reveals dislocations and cracks that are not visible in the SE image. It is
apparent that dislocation generation has occurred along the crack path. The ECCI image

reveals that dislocations lie end-on as well as parallel to the surface, and are limited to a

 

Figure 33. Crack tip in deformed specimen 673AC under different imaging conditions.
a) High magnification SE image of the crack tip region. Inserted image shows a low
magnification SE image of the secondary crack and highlighting the area analyzed.

b) ECCI image of the crack tip region showing dislocations lying end-on.

65

 

Figure 34. Examples of dislocation distribution left in the wake of the cracks for 1590FC
specimen. a) SE image of the crack edge. b) ECCI image of the same crack edge
region.

66

region within ~ 4-um of the crack. A thin bright band is observed near the edge of the
crack. One possible reason for this bright band is the result of the edge effect. Figure 35
shows a schematic representation of the interaction volume within the specimen and near
the edge of the specimen. As shown, backscattered electrons can escape easily through
the edge and hence result in a higher BSE. In contrast, backscattered electrons generated
away from a edge must escape through the surface, consequently there will be a lower
BSE yield. However, dark bands at crack edges have also been observed. This may be
rationalized by considering the possible effects of mode I and II failure, where the surface
may be tilted (Figure 36) leading to a local change in channeling conditions (this could
result in either a higher or lower BSE yield).

Figure 37 shows SE and ECCI images of an edge of the crack of 673AC
specimen, in this case, with a dark band near the crack edge. Most of the dislocations are
seen to be lying end-on. The bright region with some dark patches is indicative of the
cumulative strain field of the clusters of dislocations in the region. This suggests a third
possibility that the bright band could also be a result of numerous dislocations at the
edge, resulting in a cumulative ECCI effect [43]. A noticeable feature of the ECCI
micrograph is that more than a few microns from the crack, there are hardly any
dislocations. While propagation of a crack clearly leads to an increased dislocation
density in the local crack region, the four-point bending does not generally lead to

dislocation generation away from the crack regions.

67

Backscattered detector

<— electron beam—+

 

I BSE yield
Stronger BSE yield
BSE yield
edge of - .. 1!"!
”rpm. n. .
SPCCimen ‘6'.“7.“ 71:47, (15;; ::S-\~ SpCClmen

Q

‘\

‘9“
‘2“ ‘
—————/

:5. [ ’i‘flf 1",“ Q
interactio: / 40"”
volume

Figure 35. A higher BSE yield resulted when the rastering beam is near to a edge of a
specimen.

68

   
 
   
  

Bragg condition

slightly less than
the Bragg condition

slightly greater than
the Bragg condition

rastering electron ' '
beam

 

 

 

 

 

Figure 36. Schematic diagram showing how the combination of mode I and II tear results
in a change in the BSE yield for both sides of the crack edge.

69

 

Figure 37. Images of a crack edge of a 673AC specimen illustrating a) SE image and
b) ECCI image with dislocations in dark/bright contrast lying end-on.

70

A comparison of dislocation densities a few tens of microns away from the crack
before and after the deformation showed no significant change in the dislocation density.
As shown in Figure 38, both micrographs show the same dislocations (arrowed). As
expected, it appears that the dislocations generated in 4-point bend testing are limited to
regions very close to the cracks. The difference in the contrast of the two images is due
to differences in channeling conditions used to form the images. General observation of
the specimens shows that the toughened materials (673AC) displayed a significantly
higher dislocation density over a larger region in front of the crack tip than the brittle
materials (159OF C). This can be seen in Figure 39 where the dislocation density and
distribution in front of the crack tips of the brittle materials, 1590FC, are less than that of
the toughened material, specimen 673AC.An interesting observation using the ECCI
technique is that by changing the channeling conditions between positive (+) g and
negative (-) g subsurface defects can be distinguished as to a dislocation or microcrack.
Dislocation images will display a dark/bright contrast or a bright/dark contrast image
pairs (Figure 40) as a result of the change in the (+) or (-) channeling condition. This
change in appearance is somewhat similar to that observed between darkfield and
brightfield images commonly used in TEM. However, the actual contrast mechanisms
are quite different. Cracks on the other hand do not show changes in contrast with
changes in channeling conditions as the electron beam will penetrate deeper into the
surface and hence cracks will continue to appear dark (as in Figure 41) in all channeling

conditions.

7l

 

Figure 38. Images of an area 80 pm from the edge of the crack for 673AC specimen
before (a) and afier (b) 4-point bend test. No significant change in the dislocation
density was observed.

72

 

Figure 39. Comparison of dislocation densities for a) 673AC specimen, b) 1590FC
specimen.

73

 

Figure 40. Changes in contrast of dislocation images when the channeling band is
changed from a) + g to b) - g.

74

 

Figure 41. Cracks do not show any change in contrast as the channeling condition
changes from +g to -g.

75

3.4 Summary

The use of the electron channeling contrast imaging (ECCI) technique to

image dislocations near crack tips in bulk NiAl single crystal 4-point bend specimens has

led to the following observations:

A higher dislocation density was observed over a larger region in front of the crack tip
for the toughened materials as compared to the brittle materials.

Although a different contrast mechanism is used in ECCI, dislocations observed near
the crack tips using this technique appear somewhat similar to images of dislocations
lying normal to TEM thin foils.

A bright or dark region along crack edges and in front of crack tips is believed to be
due to the local tilt of the crystal at the crack edge/tip. This local tilt may be a result
from a number of mechanism. Dislocations originating from the crack could change
the channeling condition resulting in a higher BSE signal. Backscattered electrons
escaped though the crack edges instead escaping through the surface thus resulting in a
higher BSE yield. The same band could also be a result of the combination of mode I
and II failure causing the crack surface to be tilted such that the resultant BSE yield
from the tilted edge will be greater or less than the surrounding area, thus resulting in
the change of contrast.

It was observed that the dislocation density was essentially unchanged in regions
more than a few microns from the crack path. Consequently, it appears that the
dislocations generated by 4-point bend testing are limited to regions close to the

crack, even in toughened materials.

76

0 Changing the channeling conditions from (+) g to (-) g result in dislocation images

that are similar to brightfield/darkfield images obtained in TEM.

CHAPTER 4

DISCUSSION

In chapter 2 and 3, TEM and ECCI examinations of dislocation mobility and
generation in single crystal NiAl have been presented. In this section, these results will be
used to explain the observed mechanical behavior of these NiAl alloys.

Previous studies have clearly shown that the room temperature ductility and
toughness of NiAl are strongly influenced by controlled heat treatment, use of superimposed
hydrostatic pressure, microalloying with agents, and/or by the presence of interstitial
impurities. These studies have lead to the proposition that NiAl and BCC metals having
significant levels of interstitial impurities, are subject to a strain aging phenomenon.

It is clear from both the TEM and ECCI results that changes in thermal treatment
have an effect on the observed dislocation mobility and distribution in NiAl single crystals.
It is of interest to attempt to relate these observations to the general understanding of strain
aging in this material.

In the TEM examination, it was observed that air cooled specimens displayed lower
yield strengths but correspondingly higher fracture toughnesses in contrast to furnace cooled

specimens which tended to display higher yield strengths but were correspondingly very

77

78

brittle. Furthermore, the deformed specimens showed vast differences in the gross
dislocation morphology and densities depending on whether they had been air cooled or
furnace cooled. Furnace cooled specimens displayed dislocations that were heavily tangled,
whereas air cooled specimens exhibited fewer dislocation tangles. However, the gross
dislocation morphology and densities prior to deformation were the same irrespective of heat
treatment histories and cooling rates. Additionally, dislocations in both the deformed and
undeforrned specimens were observed to move under the influence of the electron beam
regardless of pre-deformation cooling rate. This suggests that dislocations are mobile in
NiAl regardless of thermal history. However, as the dislocations which glided under the
influence of the electron beam moved into dislocation tangles, they were found to be
restrained from moving further and piled-up with the rest of the dislocations in the tangles.

A possible explanation for the vast differences in the gross dislocation morphology of
the deformed specimens is the cooling rate used. It should be recalled that commercially
pure single crystal NiAl alloy, which has significant levels of interstitial elements, was used
in this study. It is likely that in specimens that have been slow cooled, interstitial atoms have
sufficient time gettering to diffuse into the cores of dislocations, resulting in pinning of the
existing dislocations and leading to restricted mobility. When such a specimen is then
deformed, newly generated dislocations, which themselves may not be pinned, will however
be restrained by the rest of the pinned pre-existing forest of dislocations. This leads to
increases in dislocation pile-ups, dislocation tangles and the accumulation of more forest
dislocations, resulting in increases in yield stress and consequently, a lower ductility. This is
essentially a static strain aging phenomenon as it is the dislocations which are pinned prior to

deformation which are responsible for the yield strength increase. In contrast, fast cooling

79

‘traps’ the interstitials in solution and as such pre-existing dislocations are not subject to the
static strain aging and are free to move. When these fast cooled specimens are deformed,
newly generated dislocations and pre-existing dislocations will encounter fewer obstacles,
thus resulting in a lower yield stress and higher ductility for the material.

The actual interstitial agents responsible for the strain aging phenomenon have not
been determined in the present study. However, Weaver [21] has extensively investigated the
strain aging phenomenon in NiAl. Through a systematic study using several nominally
stoichiometric polycrystalline and single crystal alloys containing varying interstitial
contents, it was concluded that the major solute responsible for pinning the dislocations in
NiAl is carbon.

In this present study, commercial purity single crystal with an average of 100 wppm
of carbon and 50 wppm of oxygen were examined [8,9,13]. The large amount of
impurities/interstitial atoms present provide a good opportunity for static strain aging to
occur. However, with a significant reduction in the concentrations of interstitial agents, as in
high purity materials, it is believed there will be less pinning of the pre-existing dislocations.
This will decrease the susceptibility of NiAl to the static strain aging phenomena as pre-
existing dislocations will remain mobile despite slow cooling. Consequently, high purity
materials should have higher ductilities and fracture toughnesses. This has been
demonstrated by Levit et a1. [19] who have achieved elongations of ~25% in high purity
materials. In a separate study employing in-situ straining TEM, Ghosh and Crimp [57]
further reinforced the concept of strain aging in commercial purity NiAl when they observed
the motion of dislocations in commercial purity single crystals NiAl to be ‘slow and

progressive’ versus very rapid in high purity materials. This slow and progressive motion of

80

purity single crystals NiAl to be ‘slow and progressive’ versus very rapid in high purity
materials. This slow and progressive motion of the dislocations in commercial purity
alloys is believed to be a consequence of solute drag (dynamic strain aging). In the high
purity single crystal NiAl, the lower interstitial contents lead to a much lower degree of
dynamic strain aging.

While TEM examinations indicate that static strain aging plays a role in the effect
of thermal treatments, ECCI images, showing decreased crack tip plasticity in brittle
materials, have not been able to specifically determine if a static strain aging phenomena
is involved, or if some other effect, such as interstitial embrittlement, is responsible for
decreased dislocation generation. Nevertheless, ECCI images provide important
information on how the effect of thermal treatment affects the crack tip plasticity.

The ECCI observations show that in the air cooled specimens (toughened) the
dislocations ahead of the crack tips were more widely distributed (~4 um ahead of the
crack tip) than furnace cooled specimens (~2 um ahead of the crack tip) (refer to Figure
39). This suggests that it is more difficult to generate fresh dislocations at the highly
stressed region of the crack tip in fumace cooled specimens. When dislocation
generation is difficult, ductility can suffer if pre-existing dislocations cannot provide
sufficient plastic deformation to relax the crack tip.

The sparse distribution of dislocations observed in fi'ont of crack tips in the
furnace cooled specimens might also another possibility. Assuming that new dislocations
can be generated at crack tips in both the air cooled and fumace cooled materials, then
differences in the pre-existing dislocations described above may result in differences in

the observed dislocation behavior at crack tips. The large number of dislocations

81

observed in front of crack tips of air cooled specimens would then suggest that newly
generated dislocations, with fewer pinned dislocations or dislocation tangles ahead of
them, are able to move further away from the crack tips and are more readily observed by
ECCI. In contrast, if dislocations are pinned very close to the crack tips, as in the furnace
cooled materials, these dislocations might fall within the bright or dark bands near the
crack tip, making the dislocations difficult to observe.

When a crack tip emits dislocations, these newly generated dislocations must
move out of the crack tip region. But with dislocation tangles resulting from solute
pinning the dislocations in furnace cooled materials, these dislocation tangles may hinder
the emission of dislocations from the crack tips as the dislocations will be block by the
dislocation tangles. As a result, fewer dislocations move out of the crack tips. This may
be what was observed in the ECCI images for furnace cooled materials. With fewer
dislocations emitting out of the crack tip, there will be less plastic deformation and crack
blunting and hence cracks continue to propagate and this results in brittle fracture.

In general, this study demonstrates that TEM and ECCI can play complementary
roles in assessing the role of dislocations in controlling the mechanical response of the
material. TEM provides the microstructural information on how the dislocation mobility
is hindered by forest of dislocations as a result of static strain aging. ECCI, on the other
hand, shows dislocation distributions ahead of the crack tip region. These dislocation
distributions are indicative of the crack tip plasticity. Relative toughness or brittleness of

a material may be estimated by noting the extent of plasticity surrounding the crack tip.

[58].

CHAPTER 5

GENERAL CONCLUSION

This study examines the types of dislocations, their behavior, as well as the

dislocation distribution in the region of the crack tips and crack edge, of single crystal

stoichiometric NiAl.

Slip trace analysis revealed ( 0 0 1) to be the glide plane for all the specimens.
TEM examination of both deformed as well as undeforrned specimens showed
that both < 0 O l > (dominant) and < O 1 1 > dislocations were present in
toughened as well as the embrittled specimens. This suggest that it is not the type
of dislocation present that result in the change in the properties of toughness or
brittleness of the material.

Irrespective of heat treatment histories and yield strengths, dislocations were
observed to move under the influence of the electron beam suggesting that the
dislocations in the structure are mobile.

TEM observations of deformed slow cooled specimens showed dislocation
tangles or forest of dislocations, while fast cooled specimens exhibited very few
dislocation tangles. Dislocations were observed to be restrained when they move

into these tangles. This suggests that the dislocation tangles are obstacles to

82

83

dislocation motion. These dislocation tangles are a result of static strain aging
where the interstitial atoms diffuse into the cores of the dislocations and pin them.
As a result, the pre-existing dislocations are immobile.

Electron channeling contrast imaging (ECCI) allows near sub-surface defects to
be imaged. Examinations of crack tips and crack edges showed that fast cooled
specimens displayed a higher dislocation density over a larger region in front of
the crack tips than the slow cooled specimens. This dislocation distribution ahead
of the crack tip is indicative of the crack tip plasticity. Furnace cooled specimens
have reduced crack tip plasticity and air cooled specimens have higher crack tip
plasticity.

Lastly, TEM and ECCI play complementary roles in assessing the role of

dislocations in controlling mechanical response of material.

10.

11.

12.

REFERENCE

. D.P. P0pe and R. Darolia, MRS Bulletin/May 1996, Vol. 21 , No. 5, pp. 30-36.

R. Darolia, JOM 43 (1991) pp. 44-49.

. J .H. Westbrook, MRS Bulletin/May 1996, Vol. 21, No. 5, pp. 26-28.

P. Nash, M.F. Singleton and J .L. Murray, Phase Diagrams of Binary Nickel Alloys, Vol.
1, ASM International, Metals Park, Ohio (1991).

. D.R. Miracle, Acta metall. mater., 41, No.3 (1993), pp. 649-684.

C.T. Liu, and J. A. Horton Jr., Materials Science and Engineering A, 192/193 (1995) pp.
170-178.

B. Zeumer, W. Wunnike-Sanders, G. Sauthoff, Materials Science and Engineering A,
192/ 193 (1995) pp. 817-823.

J .E. Hack , J .M. Brzeski and RD. Field, Scripta Met. et Mat., 27, (1992) pp. 1259-1263.

J.E. Hack , J.M. Brzeski, R. Darolia and RD. Field, High Temperature Ordered
Intermetallic Alloys V, MRS Proceedings, Vol. 288, 1993 MRS, pp. 1197-1203.

American Society for Metals, Metals Handbook, 9th Ed. Vol. 11, ASM, Metals park, OH
(1986)

D. Hull and DJ. Bacon, “Introduction to Dislocations”, 3rd Edition, International series
on Materials Science and Technology, Vol. 37, Pergarnon Press, (1984) , pp. 225.

AH. Cottrell, “Dislocations fld Plastic Flow in Crvstfi”, Oxford University Press,
Amen House, London, 1965.

 

84

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

85

J .M. Brezeski, J .E. Hack, R. Darolia and RD. Field, Materials Science and Engineering,
A170 (1993) pp. 11-18.

R.W. Margevicius and J.J. Lewandowski, Acta metall. mater., Vol. 41, No. 2, (1993) pp.
485-496.

R.W. Margevicius and J .J . Lewandowski, Scripta Met. et Mat., Vol. 25, (1991) pp. 2017-
2022.
D.S. Liu, M. Manoharan, and J .J . Lewandowski, Scripta mater., 23, (1989) pp. 253-256.

C.T. Liu, High Temperature Ordered Intermetallic Alloys V, MRS Proceedings. Vol 288,
1993 MRS, pp. 3-19.

R. Darolia, D.F. Lahrman and RF. Field, Scripta Met. et Mat., Vol. 26, (1992) pp1007-
1012.

v.1. Levit, I.A. Bul, J. Hu and M.J. Kaufman, Scritpa Met., Vol. 34, No. 12 (1996) pp.
1925-30.

W.C. Leslie, Eds., Quench and Strain Aging, Vol. 5, MIT Press, Cambridge, MA., 1986.

ML. Weaver, “Investigation of Strain Aging in the Ordered Intermetallic Commund fl-
Ni_a_l”, A Dissertation, University of Florida, 1995

 

 

S.J. Chang and SM. 0hr, Journal of Applied Physics, 52, (1981), pp. 7174-7181.

S.M. 0hr and 8.]. Chang, Journal of Applied Physics, 53, (1982), pp. 5645-5651

S.M. 0hr, Scripta Metall., Vol. 20, (1986), pp. 1501 -1505.

I.H. Lin and R. Thomson, Acta Metall., 34, (1986), pp. 187-206.

B.S. Majumdar and 8.]. Burns, Acta Metall., 29, (1981) pp. 579-588.

H.L. Fraser, R.E. Smallman, and M.H.Loretto, Phil. Mag, Vol. 28, (1973), p. 651-665.

Maria A. Morris, Jean-Francois Perez, and R. Darolia, Phil. Mag., Vol. 69, (1994), pp.
485-506.

R.D. Field, D. F. Lahrman and R. Darolia, Acta metall. mater., Vol. 39, No. 12, (1991),
pp. 2951-2959.

M.A. Morris, J .F. Perez and R. Darolia, High Temperature Ordered Intermetallic Alloys
VI, MRS Proceedings Vol. 364, 1995 MRS, pp. 413-418.

31.

32.

33

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

86

RD. Field, D.F.Lahrman and R. Darolia, High Temperature Ordered Intermetallic Alloys
IV, MRS. Proceedings Vol. 213, 1991 MRS, pp. 255-260.

R.D. Field, D.F.Lahrman and R. Darolia, Acta metall. mater., Vol. 39, No. 12, (1991) pp.
2951-2959.

. M.H. Loretto and R. J. Wasilewski, Phil. Mag, 23, (1971) pp. 1311-1328.

R.D. Field, D.F. Lahrman and R.Darolia, Acta. metall. mater., Vol. 39, No.12, (1991) pp.
2961-2969.

D.B. Miracle, S. Russell and CC. Law, High Temperature Ordered Intermetallic Alloys
III, MRS Proceeding, Vol 133, 1989 MRS, pp. 225-230.

M.J. Mills and DB. Miracle, Acta metall. mater., Vol. 41, No. l (1993) pp. 85-95.
H.L.Fraser, M.H. Loretto and R. E. Smallman, Phil. Mag., Vol. 28, (1973) pp. 667-677.

D. Hull and DJ. Bacon, Introduction to Dislocations, 3rd Edition, International series on
Materials Science and Technology, Vol. 37, Pergamon Press, (1984) , pp. 35.

AK. Head, P. Humble, L.M. Clarebrough, A.J. Morton and CT. Forwood, Defect in
Cgstalline Solids, Vol. 7, ed. S. Amelinckx, North-Holland Pub. Co. ( 1973).

J.T. Czernuszka, N.J. Long and RB. Hirsch, Inst. Phys. Conf,, Ser. No 117: Section 10
(1991)

DC. Joy, MRS Proceeding, Vol 183, Materials Research Society, (1990), pp. 199-209.
A.J. Wilkinson and PB. Hirsch, Phil. Mag., A, 72, No. l (1995), pp. 81-103.

A.J. Wilkinson, G.R. Anstis, J.T. Czernuszka, N.J. Long and RB. Hirsch, Phil. Mag. A, 68
(1993), pp. 59-80.

B.-C. Ng, B.A. Simkin and MA. Crimp, Deformation and Fracture of Ordered Intermetallic
Materials 111, edited by W.O. Soboyejo, T.S. Srivatsan and H.L. Fraser, The Minerals,
Metals and Materials Society, (1996), pp. 337-348.

S.M.Ohr, Materials Science and Engineering, 72 (1985), pp. 1-35.

1. Baker, and J .A. Horton, Phil. Mag. A, 67, No 2 (1993), pp. 479-489.

 

87

47. B. Ghosh and MA. Crimp, Deformation and Fracture of Ordered Intermetallic Materials
m, edited by W.O. Soboyejo, T.S. Srivatsan and H.L. Fraser, The Minerals, Metals and
Materials Society, (1996), pp. 349-359.

48. M. Goken, H. Vehoff, and P. Neumann, Scripta Met. et Mat,, 33 (1995), pp. 1187-1192.

49. T. Ichinokawa, M. Nishimura, H. Wada, Journal of the Physical Society of Japan, 36 No. l
(1974), pp. 221-226.

50. D.G. Coates, Phil. Mag, 16 (1967), pp. 1179-1184.

51. G.R.Booker, A.M.B. Shaw, M.J. Whelan, and RB. Hirsch, Phil. Mag, 16 (1967), pp.
1185-1191.

52. BA. Simkin, M.A. Crimp, Proceedings, Microscopy and Microanalysis, edited by G.W.
Bailey, J .M. Corbett, R.V.W. Dirnlich, J .R. Michael and NJ. Zaluzec, 1996, pp. 140-141.

53. J.T. Czemuszka, N.J. Long, E.D. Boyes, and RB. Hirsch, MRS Proceeding, Vol. 209
(1991), pp. 289-292.

54. RB. Hirsch, Electron Microscopy of Thin Crvstefi, Robert E. Krieger Publishing Co.,
Inc., Hungtinton, New York, (1977), pp. 274.

55. N.Rusovic and ET. Heuing, Phys. Stat. Sol. A, 57, (1980), pp. 529.

56. J. Goldstein, Practical Scanning Electron Microscopy; Electron and Ion Microprobe
Analysis, New York, Plenum Press, 1975.

57. B. Ghosh and MA. Crimp, “TEM in-situ straining and study of dislocations in NiAl”, to
be published.

58. R.W. Hertzberg, “Deformation and Frgcture Mechanics of Engineering Materials”, 3rd Ed,
John Wiley and Sons, 1976.

28

fracture toughness, whereas fast cooling (air cooled) tends to decrease the yield strength
but results in a much higher fracture toughness value. The dislocation structures of each
specimen were examined and the Burgers vectors b were determined using the g°b=0
invisibility criterion. Generally dislocations will be invisible when their b lies in the
reflecting plane. Using two-beam conditions for a series of individual reflections, two
such conditions of invisibility should be found. As b is common to both reflections, b

must be the zone axis of the two planes.

2.4.1 Undeformed Specimens

Burgers vector characterization of an undeforrned 673F C specimen is shown in
Figure 12. The “a” dislocations with b = [ 0 0 1 ] are observed to be in weak contrast for
both [2 0 0 ] and [ 1 1 0 ] reflections while the “b” dislocations b= [ 0 T 1 ] are observed
to be in weak contrast in both [2 0 0 ] and [ 0 1 1 ] reflections.

Figure 13 shows micrographs of undeformed 473FC and 673AC specimens
exhibiting <001> and <01 l> dislocations. Likewise, undeforrned 1590AC and 1590F C
specimens show similar types of dislocations. It is noted that majority of the dislocations
displayed in all the undeforrned specimens are dominantly <001> type with a few
scattered <011> dislocations.

The gross morphologies of all the undeforrned specimens show similar dislocation
densities. Figure 14 shows micrographs of 473FC, 673F C and 673AC specimens

showing single dislocations widely spaced with few dislocation loops and tangles.

 

 

 

' 5‘4-

F1
.

Will

U'S

If.

llllllllllllllllf