Khan-1111*" '
1‘ ' ' q

. - ”‘4'? -H" A

“7 MA

 

1
l

u—w

 

, 23-»:
‘w-l—VO:

 

M53

”200%
573l ft

17

This is to certify that the
dissertation entitled

DISLOCATION GENERATION AND MOTION IN B2 NIAL
SINGLE CRYSTALS

presented by

BIMAN GHOSH

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemical Engineering and
Materials Science

 

 

WZ/é {—3

Major Professor’s Signgu

December 1 8, 2003

 

Date

(OUIF“i“*“ [-JA hrjA "WA Y
L-llilerli fl

MAT ERIAL in 88“”

 

PIACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FlNES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/Date0ue.p65-p. 15

DlSLOCATl‘

 

DISLOCATION GENERATION AND MOTION IN I32 NIAL SINGLE CRYSTALS
By

Biman Ghosh

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemical Engineering and Materials Science

2003

 

DISLOCATIO‘.

in Order I
dislocation bcha
studied followin
electron micros.
dislocation subs
changes in the C
observed in all t
generated dileC
n1 edge and f’”
in the high PL”

mmmCrClally pU

 

incontrast 10 ll:
Uniform. The “T
explained in ten‘
henna] history
commercially p‘u
in-situ straining.

In the high purir

 

ABSTRACT
DISLOCATION GENERATION AND MOTION IN 82 NIAL SINGLE CRYSTALS
By

Biman Ghosh

In order to understand the effects of composition, orientation, and cooling rate on
dislocation behavior in MA], high and commercial purity NiAl single crystals have been
studied following thermal treatments, deformations, and in-situ straining in transmission
electron microscopy. It was observed that the yield strength and post deformation
dislocation substructure of the commercial purity alloy were impacted more by the
changes in the cooling rate than the high purity alloy. Dislocation motion was readily
observed in all the samples. The pre-existing dislocations were sessile, while the newly
.generated dislocations glided in slip bands. Dislocations were generated both from the
foil edge and from the foil interior. Cross-slip of mobile dislocations was often observed
in the high purity alloy whereas slip was found to be much more planar in the
commercially pure alloy. Dislocations moved by rapid jumps in the high purity material,
in contrast to the commercially pure material where movement was much slower and
uniform. The differences in the mechanical properties and dislocation behaviors are
explained in terms of static and dynamic strain aging. Despite the well-known effect of
thermal history on the mechanical behavior of MA], variations in cooling rate of the
commercially pure material did not appear to influence the dislocation behavior during
in-situ straining. In all cases <001> and <110> dislocations were observed to be mobile.

In the high purity material mobility of <1 1 1> dislocations was also observed.

 

Dedicated to my parents, Mrs. Anjali and Mr. Benoy Krishna Ghosh.

iii

 

lwould '
|
illllltllll whose g
complete this it I

Special n
EsinInce. and r
Thomas Bicler.
mggtstions and
Naval Research
gratefully ackno.

Would ;
EVandeilc, 0H
Willing lhe N
Peluolo and the
graduate SIUde
pmgram.

l WOuld
encouragemcm a
and wife‘ Sufita

i ,
aways there b\‘ I

 

 

 

ACKNOWLEDGEMENTS

I would like to thank everyone who stood by my side in this entire journey and
without whose guidance, help, and support it was not possible for me to stay focused and
complete this work.

Special mention to my mentor and advisor, Dr. Martin A. Crimp, for his guidance,
assistance, and painstaking reviews to finish this dissertation. I would like to thank Dr.
Thomas Bieler, Dr. Karen Klomparens, and Dr. David Grummon for their helpful
suggestions and advice during this investigation. The financial support of the Office of
Naval Research (Scientific Officer, Dr. George Yoder; Grant # 00014-94-1-0204) is
gratefully acknowledged.

I would also like to thank Dr. Ram Darolia at General Electric Aircraft Engines,
Evandale, OH and Dr. Vladimir Levit, University of Florida, Gainesville, FL for
providing the MA] single crystals used in this study. The technical assistance of Tom
Pelazolo and the late Leo Zafransky is also acknowledged. Special thanks to my fellow
graduate student, Boon-Chai Ng, for his assistance in running the image-simulation
program.

I would like to thank my parents, Anjali and Benoy Krishna Ghosh, for their
encouragement and support during this study. Lastly, I am grateful to my daughter, Bitan
and wife, Surita for their help, support, and persistence in this long journey and to be

always there by my side.

 

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vi
LIST OF FIGURES ................................................................................ vii
1.0 INTRODUCTION .............................................................................. 1
2.0 BACKGROUND ................................................................................ 4
2.1 Physical Properties ..................................................................... 4

2.2 Mechanical Properties ................................................................ 10

2.3 Diffusion ............................................................................... 27

2.4 Creep .................................................................................... 29

2.5 EnvironmentalResistance ............................................................ 30

3.0 EXPERIMENTAL PROCEDURE .......................................................... 34
3.1 Thermal Treatments .................................................................. 36

3.2 Compression Sample Preparation ................................................... 38

3.3 Compression Deformation ........................................................... 38

3.4 TEM Sample Preparation ............................................................ 40

3.5 Image Simulation ..................................................................... 41

3.6 In-situ Tensile Straining .............................................................. 42

3.7 In-situ Tensile Deformation of Disc Samples ..................................... 45

4.0 EXPERIMENTAL RESULTS ............................................................... 48
4.1 Mechanical Behavior ................................................................. 49

4.2. Slip Trace Analysis .................................................................. 53

4.3 Static/Post-deformation TEM Observations ....................................... 53

.4.4 Image Simulation ..................................................................... 71

4.5 In-situ Transmission Electron Microscopy ........................................ 80

5.0 DISCUSSION ................................................................................. 136
5.1 Structure—Property Correlations ................................................... 136

5.2 Static Strain Aging .................................................................. 140

5.3 In-situ Straining TEM and Dislocation Behavior ................................ 148

5.4 Dynamic Strain Aging ............................................................... 159

6.0 CONCLUSION ............................................................................... 175
APPENDIX ......................................................................................... 17 8
REFERENCES .................................................................................... 1 79

 

TABLE 1. Solid

 

|
Table ll. lmpuri‘

Table III. C omp

Table IV. Yield I
“Isiah. .......

 

Table V. Burg.

Table \’l. Paran

Table VII. DlSlt

 

 

Tible VIIL Con
“El“ cYE’Stles . ,

LIST OF TABLES

TABLE 1. Solid Solution Strengthening in NiAl [7] .......................................... 19
Table II. Impurity Concentration of the MA] Single Crystals ................................ 35
Table III. Comparison of yield stress obtained by compression test of commercially pure
NiAl single crystals with fracture toughness. ................................................... 50
Table IV. Yield stress properties obtained by compression test of high purity NiAl single
crystals. ..... . ......................................................................................... 52 _
Table V. Burgers vectors analyses for deformed specimens in relation to their thermal
histories and yield stress in commercial purity materials ..................................... 70
Table VI. Parameters used to simulate dislocation ‘A’ ....................................... 79

Table VII. Dislocation velocities measured at the onset of straining using CCTV camera
......................................................................................................... 85

Table VIII. Comparison of calculated and measured [138] activation volume for NiAl
single crystals ...................................................................................... 170

vi

Figure 1. Bl std
B) occupy the CI

Figure 2. Phasl
the Ni-Al alloy I

Figure 3. Lattice

 

Figure 4. Figure
Dislocation line

Figure 5. Yield :
Figure 6. Relati.

Figure 7. Roo:
IIIICI’OdllOfing a

Figure 8. Dim
Species in NiAl

Figure 9 SCIIC‘
Hack er al. [13,

Figure 10. on-

u..
'c
‘..
I

LIST OF FIGURES

Figure 1. BZ structure, also known as the ordered bcc lattice. Two types of atoms (A and
B) oCcupy the corner positions or body center position respectively. ........................ 5

Figure 2. Phase diagram illustrating the phase stability as a function of temperature for
the Ni-Al alloy system. .............................................................................. 7

Figure 3. Lattice parameter and density of NiAI vary as a function of stoichiometry. .....9

. Figure 4. Figure 4. Dislocation line energy in stoichiometric NiAl at room temperature.

Dislocation line directions are measured from [001] direction [41]. ........................ 13
Figure 5. Yield stress of NiAl varies strongly as a function of stoichiometry ............. 18
Figure 6. Relationship between hardening rate and size of the solute element ............ 20

Figure 7. Room temperature ductility of <110> oriented NiAl single crystal with

microalloying additions. ........................................................................... 23
Figure 8. Diffusion coefficients (a) and activation energy (b) for diffusion of various
species in NiAI as a function of stoichiometry ................. . ................................ 28
Figure 9. Schematic representation of the thermal treatment process in accordance with
Hack et al. [18, 19] ................................................................................. 37
Figure 10. Orientation of the specimen .within the stereographic triangle .................. 39
Figure 11. Schematic of the 15 mm long and 5 mm wide in-situ tensile specimen ...... 43

. Figure 12. Change of yield stress and fracture toughness (Per Hack et al. [18, 19]) with
thermal history. The inverse relationship of yield stress and fracture toughness may be
noted. .................................................................................................. 51

Figure 13. Two faces of a deformed specimen showing slip traces with measured angles
a and B on (a) front and (b) right faces respectively. .......................................... 54

Figure 14. Schematic representation of compression sample showing the measurement of
a and [3 used in determining the slip plane. ..................................................... 55

Figure 15. Stereographic projection of the single crystal compression sample shown in
figure 13 displaying the plotting of angles a and [3 used to determine the slip plane. . . ..56

vii

their A“
more of co.
angles in (a) an .
(cianundclomi

Figure 17. The

nr§=min.

Figure 18. Hca
FC specimen; (a

 

 

Figure 19. Dct‘t‘
lalbrighttield

Figure 20. Det‘t
dislocations in (

Figure 21. C on
of dislocations.

diierent operat
6150 56611 [0 m0

HWWZL Di
SF‘E‘imens; (a)

”dell VA
673 AC Specin

Tlgul'e 24 l

dislocations A

“Web.
dlSlOQatiOnS [I
cmldillbn

. a (
WHI_ pp
“Role .
a) in d

lam h

ealll'] g

Figure 16. Isolated, widely spaced, single dislocations are observed throughout the
structure of commercially pure NiAl single crystals with few dislocation loops and
tangles in (a) an undeformed 473 PC specimen. (b) an undeformed 673 PC specimen and
(c) an undeformed 673 AC specimen respectively. ............................................ 57

. Figure 17. The same area of undeformed commercially pure 673 PC specimen is shown
with E: [011], [200] and [110]; ‘a’ denotes [001] and ‘b’ denotes [0T 1] slip directions.
......................................................................................................... 59

Figure 18. Heavily tangled dislocation morphology in deformed commercially pure 473
PC specimen; (a) brightfield image and (b) weak beam image. ............................ 61

Figure 19. Deformed commercially pure 673 FC specimens reveal tangled dislocations;
(a) brightfield and (b) weak beam image. ....................................................... 62

Figure 20. Deformed commercially pure 1600 PC specimens display heavily tangled
dislocations in (a) brightfield and (b) weak beam image. ..................................... 63

Figure 21. Commercially pure 673 FC specimen reveals presence of two different types
of dislocations, a with b = [100] and b with b = [01I]. The same area is shown with

different operative reflections; g: (a) [I 10], (b) [01 I ] and (c) [020]. Dislocations were
also seen to move in (c). ........................................................................... 65

Figure 22. Dislocation structures observed in deformed commercially pure 1600 AC
specimens; (a) brightfield image and (b) weak beam image. ................................. 66

Figure 23. Very few dislocation tangles were observed in deformed commercially pure
673 AC specimens; (a) bright field image and (b) weak beam image. ...................... 67
Figure 24. Commercially pure defonned 673 AC specimen shows ‘a’ and ‘b’
dislocations with b = [T00] and [01T] respectively. The same region with different
operative reflections are shown in (a) [I TO], (b) [200] and (c) [2T I]. ..................... 68

Figure 25. Dislocation analyses in high purity 1600 AC specimen shows [001]
dislocations in different reflections. Dislocations that are mobile under beam heating
condition appear with good contrast in figure (a) and out of contrast in figures (b), (c) and
(d). ................................................................................................ 72

Figure 26. High purity I600 FC specimen shows [001] mobile dislocations (marked as
a) in different reflections. Type D dislocations, which are not moving due to the effect of
beam heating, also seem to have similar Burgers Vectors. .................................... 73

viii

 

Figure 2]. Dis
dislocations in l
direction in fig:
present in the n
and (c). .........

Figure 33' h l“
show such dish
rm Willi 311d

Figure 39. Drsl
for image simu'.

Figure 30. A
diffraction COIILi
ditiraction cond

Figure31. Dis‘.

olerents capturI
aresult of strai:I

Figure 31 (conti
and cross-past r

(Cl and (f) reveai

ondislocations.

Figure 3] (cont
hands that prod
Slay sessile thros

Figure 32. A r
eIllibil tangles .
Te“ (“Files are ..

Figure 33. (a) I
do of the foij
{’38th Showinv
Itiils indicatinrfl

hteen slip bail

Figure 34‘ Dis
Thlmobiledislii
tile foil and J
Sl-iHraces left ii"

i:

 

Figure 27. Dislocation characterization in high purity 673 AC specimens shows [101]
dislocations in different reflections. Dislocations, which are almost parallel to the beam
direction in figure (a), are characterized with reference to several immobile dislocations
present in the microstructure. Such mobile dislocations are out of contrast in figures (b)
and (c). ................................................................................................ 74

Figure 28. It is not usual to find [111] mobile dislocations in NiAl. Figures (a) and (c)
show such dislocations in good contrast whereas mobile dislocations are observed with
very weak and residual contrast in figures (b) and (d) in high purity 673 FC specimens.
.......................................................................................................... 75

Figure 29. Dislocation ‘A’ in an undeformed commercially pure 673 AC specimen used
for image simulation and generation of computed image. .................................... 77

Figure 30. A set of experimental images (a) - (g) of dislocation ‘A’ taken with the
diffraction conditions given in Table IV and four sets of computed images for the same
diffraction conditions corresponding to the Burgers vectors shown. ........................ 78

Figure 31. Dislocation motion in high purity fast-cooled [110] oriented NiAI. Sequence
of events captured as (a), (b) and (c) revealed progressive movement of dislocations. As

a result of straining dislocations generated within the foil and starts moving towards the
edge. .................................................................................................. 81

Figure 31 (contd.). Dislocations, almost perpendicular to the foil moved in the slip bands
and cross-past pre-existing dislocations in the plane of the foil. Sequence of events (d),
(e) and (f) revealed Slight waviness in the slip bands that indicates cross-slip of such end-
on dislocations. ..................................................................................... 82

Figure 31 (contd.). Eventual crack initiation from the edge of the foil along the slip
bands'that produces avalanche of dislocations at its tip but the pre-existing dislocations
stay sessile throughout the process. .............................................................. 83

Figure 32. A representative microstructure of commercially pure NiAl Single crystal
exhibit tangles of dislocations. Many of the dislocations are extensively jogged and a
few dipoles are also observed. .................................................................... 87

Figure 33. (a) Bright field image illustrating the nucleation of mobile dislocations at the
' edge of the foil in commercially pure air-cooled NiAI oriented to [238]. (b) and (c)
lrnages showing the dislocation motion in two slip bands displaying straight dislocation
trails indicating no cross-slip. (d) Eventual cracking from the foil edge, propagating
between slip bands. ................................................................................. 88

Figure 34. Dislocation motion in commercially pure fast-cooled [001] oriented NiAl.
The mobile dislocations are referenced with the pre—existing dislocation lying in the plane
of the foil and seen at the left side of the micrographs. Planar slip is evidenced with the
slip-traces lefi in the wake of dislocations. ...................................................... 90

 

Figure 35. 1'
commercially pI
dislocation trail ‘

Figure 36. L11
oriented NiAl; 1'

Figure 37. Fri
areas of the [01 I
straining, dislo
newly generate
easelb) and (cl,

 

 

 

 

  
  
   

Figure 38. Dis
dislocations are
and seen at the

Figure 39. Se
Oriented NiAl,

Tiara 40. L.
W] Oriented ,

Figure 35. Images of dislocations generated via in-situ straining of air-cooled
commercially pure single crystal oriented to [001] illustrating (a) both wavy and straight
dislocatiOn trails and (b) extensive accumulation of dislocation debris following straining.
........................................ 91

Figure 36. Limited cross-slip was observed in commercially pure fast cooled [001]
oriented NiAl; loops and debris formed over time with straining. ........................... 92

Figure 37. Prior to deformation, heavily tangled dislocations were observed in many
areas of the [001] orientated commercially pure fumace—cooled material. At the onset of
straining, dislocations channeled through the tangles (a). With continued deformation,
newly generated dislocations slipped through the pre-existing dislocations with relative
case (b) and (c). ...................................................................................... 94

Figure 38. Dislocation motion in commercially pure [001] oriented NiAl. The mobile
dislocations are referenced with the pre-existing dislocation lying in the plane of the foil
and seen at the left side of all the micrographs. Planar slip is evidenced with the slip- ,
traces left in the wake of dislocations. Most of the mobile dislocations seen in (b) were
not present in (a) and left the field of View in (c). ............................................. 95

Figure 39. Sequence of dislocation motion in commercially pure slow cooled [001].
oriented NiAl. Dislocations (marked with numbers) lying in the plane of the foil move in
a slow and uniform manner. ....................................................................... 96

Figure 40. Loops and debris formation is evident in slow-cooled commercially pure
' [001] oriented NiAI. Dislocation multiplication by operation of Single-ended Frank-Read
sources takes place. ................................................................................. 97

Figure 41. Commercially pure fast-cooled [110] oriented NiAl. (a) Dislocations slip in
the marked direction in slip hands by bowing out from the dislocation ends at the foil
surface. (b) A dark-field image showing a few mobile dislocations (arrowed) and planar
arrays of pre-existing dislocations. (c) Dislocation multiplication is taking place. ((1)
Loops and debris are also evident. ............................................................... 99

Figure 42. Series of images in commercially pure fast-cooled [110] oriented NiAl
characterizing the initial stages of dislocation motion. Figures (0) and ((1) Show
subsequent slip in well-defined narrow slip bands. .......................................... 100

Figure 43. Dislocation motion in commercially pure fast-cooled [110] oriented NiAI.
Sequence of image (a) through ((1) reveals dislocation mobility with planar slip and
relative difference in the velocity of the dislocations. ....................................... 102

Figure 44. Sequence of dislocation motion in commercially pure slow-cooled [110]
oriented NiAl. Dislocations marked 1-5 display the relative motion of the mobile
dislocations. Note the very planar slip confined in narrow slip bands. Dislocations '
marked in figure (b) appear doubled as a consequence of portions of dislocations moving
during exposure of film ............... . ............................................................. 103

Figure 45. C01
dislocations be
dislocations shi

Figure 46. Mul
sources in con
dynamically w.-
thc applied stre
to expand agar:
regretted and A

  
 
 

Figure 47. Pre~
mostly isolated
dipoles. (a) A

density compar
concentration o

[cums Dis?
Wed material
dislocation geni
W mostly plar
Fm STIOWS a

Figure 49. DA]
mill Purity [00:
endnn dislocati
the Plane of the

”g Captured

Figureso. DA]
and to move c

Figure 5] HI‘-"
Till.“ the frail»;
{Tl ”8 ”Climber

 

Figure 45. Commercially pure slow-cooled [110] oriented NiAl. With increasing strain
dislocations became tangled and loops and debris accumulated. Bowing out of the
dislocations Show that the motion was not totally inhibited by the debris. ............... 104

Figure 46. Multiplication of dislocations through operation of single-ended Frank-Read
sources in commercially pure slow-cooled [110] oriented NiAl. Images captured
dynamically with a video system showed the arrowed dislocation loop expanding under
the applied stress (a-c). The next image (d) shows that the loop has pinched-off and is set
to expand again while a neighboring loop has also expanded. (Quality of the picture is
regretted and Adobe PhotoShop was used to enhance the contrast of a dynamic image)
........................................................................................................ 106

Figure 47. Pre-existing dislocation morphology in high purity slow-cooled NiAl showed
mostly isolated dislocations; a few loops were also seen along with some pinched off
dipoles. (a) A typical area of [110] oriented single crystal with much lower dislocation
density compared to commercial purity specimens. (b) An area with unusually higher
concentration of dislocations in a slow-cooled [001] oriented specimen before straining.

........................................................................................................ 107

Figure 48. Dislocation motion at the onset of straining in high purity [001] oriented fast-
cooled material. Micrographs were captured at two different parts of the foil when
dislocation generation and movement were noticed from the edge of the foil. While slip
was mostly planar and debris were not readily formed in one part of the foil (a), the other
part shows a significant amount of dipole debris in well-defined slip bands (b).

........................................................................................................ 110,

Figure 49. Dislocations normal to the plane of the foil moved in orthogonal directions in
high purity [001] oriented fast-cooled material. The contrast from the slip traces of such
end-on dislocations was not as sharp as observed for dislocations lying at some angle to
‘ the plane of the foil. The disturbance created on the oxide layer was less pronounced for
being captured as well-defined slip traces during the exposure time of still photographs.
........................................................................................................ 111

Figure 50. Dislocations that are lying in the plane of the foil (marked 1, 2, 3, and 4) were
found to move quite rapidly in high purity [001] oriented fast-cooled material. .. .......112

Figure 51. High purity slow-cooled [001] oriented NiAl. (a) Dislocations generated
‘ within the foil and moved towards the edge. (b) The arrowed-dislocation cross-slipped
past its neighbors. ................................................................................. 113

Figure 52. Slow movement rate noticed with dislocations generated in high purity slow-

cooled [001] oriented NiAl. Dislocations moved at such a low rate that movement of
individual segments of dislocations were apparent. .......................................... 114

xi

...JN.

Figure 53. :3)
onenwd 1‘ Al
dislocatior not

Figure 54. D1
high puritj Sll‘
mobile dis teat
and orthog ina‘;
intheplar ot‘
llltltllltd‘ eufi

Figure 55. Hig
morpholo] .‘ shI
(bl Disloc tior:
existing di loci

Figareio. EM
(at Pre-er stin,
generated LlSli‘
trails. (b) lipl

Figure 57. lm.
lalDlSlOC lionl
dIFCCIIOIl fm
deionnatii tar
dislocatior sing

Figure 58. in:
[110] one ted
'd not in hit
Mlspe‘ me:

Figure 59. \r';
D3Illlslllll fl‘li;

Flgm’e 60, MI
more dis AA!
The foil. ] lSliwl
anthem” [in]

 

Figure 53. (a) Dislocations generated within the foil in high purity slow-cooled [001]
oriented NiAl. (b) Mobile dislocations developed loops and debris as a result of
dislocation motion and interactions. ............................................................ 116

Figure 54. Dislocations moving in the foil leaving well-defined trails of planar slip in
high purity slow-cooled [001] oriented NiAl (figures a and b). Line directions of the
mobile dislocations were close to the foil normal. Debris were developed at a slower rate
and orthogonal slip takes place. Figures (c) and (d) reveal orthogonal dislocations lying
in the plane of the foil at the alter in-situ straining. The dislocation density decreased as
the foil edge was approached. ................................................................... 117

Figure 55. High purity fast-cooled [110] oriented NiAl. (a) The pre-existing dislocation
morphology shows high number of nearly straight, isolated dislocations with a few loops.
(b) Dislocations moved at a very rapid rate with straining and have Slip past all pre-

existing dislocations. Cross-slip of the dislocations is evident from the wavy slip trails.
....................................................................................................... 118

Figure 56. Evolution of dislocation slip in high purity [110] oriented slow-cooled NiAl.
(a) Pre-existing dislocations lying in the plane of the foil did not move while newly
generated dislocations (lying normal to the plane of the foil) glided leaving visible slip
trails. (b) Slip bands were more evident after further deformation. ........................ 119

Figure 57. Images of dislocation motion in high purity slow-cooled [110] oriented NiAl.
(a) Dislocations generated at nearby foil edge gliding into thicker areas of the foil. The
direction of motion is marked on the figure (b) that was different than the tensile
deformation axis (TA) marked on figure (a). (b) Wavy slip trails left in the wake of the
dislocations indicate a variation in the Slip planes of the slipping dislocations. .......... 121

Figure 58. With increasing deformation, dislocation tangles developed in the high purity
[110] oriented slow-cooled NiAl. Unlike commercial purity material dislocation tangles
did not inhibit movement of the mobile dislocations (marked md in figure 64 a) in high
purity specimens. ...................... . ........................................................... 122

Figure 59. With increasing deformation, loops and dipole debris developed in the slip
bands in the highpurity [110] oriented slow-cooled NiAl. ................................. 123

Figure 60. Mobile dislocation characterization in high purity slow-cooled NiAl. The
mobile dislocations are referenced with the pre-existing dislocation lying in the plane'of
the foil. Dislocations with good contrast in (a) are observed with weak-contrast in (b)
and become invisible in (c) suggesting [001] Burgers vector. ............................... 125

xii

Figure 61. DI
dislocanons in
contrast for ti;
Burgers vector
slipbands in r
Figureslal and

Figure 62. MO
inrisibility. Ills
mobile disloca

[lOT] Burgers

the slip band n‘
existing disloc'
figures (at and

 

 

Figure 63. Prc
THEFT Willy fasr

Were observed

Figure 64. M.
Dislocations, ti
and Wire obser

FlEllie 65. m
m the fight sid
gOOd Contrasr

“This in (a i

FlguIe 66 1mLi
bands A 5121'
dlSTOQatiOH 10:)
ands are in C‘
moblllll 0f [1 j

e dislocgm0r
Sugges‘lng ”10F

 

Figure 61. Dislocation mobility was observed within a narrow slip band. Mobile
dislocations were visible in good contrast in figures (a) and (c) and appeared with weak
contrast for figures (b) and (d). These dislocations were characterized to have [100]
Burgers vector. Likewise pre-existing dislocations (marked 1 and 2 on both sides of the
slip-bands in d) are seemed to have [010] Burgers vector for becoming invisible in
figures (a) and (c). ................................................................................. 126

Figure 62. Mobile dislocations in high purity slow-cooled NiAl. In absence of complete
invisibility, the Burgers vector could not be conclusively determined. Two types of
mobile dislocations were present. The ones somewhat confined in the slip bands have

[IOT] Burgers vectors from figures (b) and (c) whereas dislocations to the right side of

the slip band may be of [1 l I] type from figure (c) and (d). The line direction of the pre-
existing dislocations lying in the plane of the foil was determined to be [010] from
figures (a) and (c). ................................................................................. 127

Figure 63. Pre-existing dislocations lying in the plane of the foil are used as reference in
high purity fast-cooled single crystal. Dislocations seen in good contrast in (a) and (c)

were observed with weak contrast in (b) and (d) and determined to have [II T] Burgers
. vector. ............................................................................................... 129

Figure 64. Mobile dislocation characterization in high purity fast-cooled single crystal.
Dislocations, those were seen in good contrast in (a) and in (b) became invisible in (c)
and were observed with weak-contrast in ((1) suggesting [100] Burgers vector. ..........130

Figure 65. The mobile dislocations are referenced with the pre-existing dislocation seen
in the right side of the micrographs high purity Slow-cooled NiAl. Dislocations seen in
good contrast in (a) and in (b) were observed with weak-contrast in (c) and became

invisible in (d) implying [.10I ] Burgers vector. ............................................... 131

Figure 66. Images from slow-cooled high purity NiAl shows mobile dislocations in slip
bands. A significant amount of deformation substructure has been formed and some
dislocation loops are also observed in the substructure. The dislocations within the slip
bands are in extinction in figure (c) and in very weak contrast in figure (b) indicating
mobility of [1 1 1] dislocations. ....... . ........................................................... 132

Figure 67. Dislocations moving in narrow slip bands in slow-cooled high purity NiAl.
The dislocations are in extinctiOn in figure (c) and in very weak contrast in figure ((1)
suggesting mobile dislocations [101] mobile dislocations. ................................. 133

Figure 68. Images from slow-cooled high purity NiAI Show mobile dislocations in slip
bands. The mobile dislocations were in extinction in figure (c) and in very weak contrast
in figure (d) suggesting [101] mobile dislocations. Similarly, some of the pre-existing
dislocations or loops may be of [010] type for being in extinction and weak contrast in
figures (b) and (c) respectively. .................................................................. 135

xiii

 

reheated to 67:
of the same ma

Figure 70. An 1..
in Figure 53 (at

Figure 71. Low
NiAl. Figure (
forming loops ;

Figure 72. Sch

Figure 73. PloI
elevated tempt
activation ener:
Microsoft EXCL

 

Figure 69. Voids observed in the substructure of high purity [110] oriented material-
reljezlted to 673 K following homogenization anneal of 1600 K. Three different samples
of the same material revealing isolated and cluster of voids at different magnifications.

-.- ..................................................................................................... 145

Figure 70. An illustration of measuring the radius of curvature for bowed-out dislocations
in figure 53 (a) using the Scion program [120]. ................................................. 155

Figure 71. Loop formation at its different stages in high purity slow-cooled [001] oriented
NiAl . Figure (a) shows dislocations separated by jogs are coming close together to start
fanning loops and figure (b) shows progression and finally pinched-Off loops. ......... 157

Figure 72. Schematic view of loop and dipole formation after Shi et al. [120]. . . .. ...158
Figure 73. Plot of diffusivity with temperature after Weaver [80] and extrapolation of

elevated temperature diffusivity values to 300 K. The frequency factor, Do and the

activation energy, Q were calculated from the equation of the straight line obtained using
Micro soft Excel. .................................................................................. 163

xiv

The on
over the last tie I
of gas turbine
presently used
further increas
above and or d'
usociated with
than stoichiom

DCVEFU;
blades and ya
united with
dei'elOpmem p
0lb’vf’dting Temp
NW Possesscg
mflductiviry‘ a:

Althoug
name”: POOr d; I
TeSlSIaDCC al ele
aSharp dque
tfinpetarure Va:

 

1.0 INTRODUCTION

The ordered intennetallic compound NiAl has received a great deal of attention

over the last few years as a potential candidate material for application in the hot sections
of gas turbine engines [1-13]. Nickel-based super alloys with Ni3Al precipitates are
presently used in the hot sections of engines for aircraft-propulsion systems. However,
further increases in engine performance and efficiency require operating temperatures
above and/or densities below those offered by the present alloys. The major limitation
associated with the present alloy system is the lower melting point and the higher density
than stoichiometric NiAl intermetallics.

Development of the next generation of high-temperature, high-pressure turbine

Nada and vanes can only be accomplished by advanced materials and processing
coupled with innovative component design. The objectives of alternate material
deVelopment programs are to have efficient blade cooling flow while increasing the
Operating temperature in conjunction with the reduction of over all engine weight [14].
NiAl possesses requisite combination of properties such as low density, high thermal
Conductivity, and excellent oxidation resistance.

Although NiAl has many useful properties, it suffers from two major drawbacks,
namely, poor ductility and toughness at ambient temperatures, and low strength and creep
resistance at elevated temperatures, particularly above 1273 K [15, 16]. NiAl undergoes
a sharp ductile to brittle transition between 673 K and 473 K, with the actual transition
temperature varying strongly with processing history, alloy composition, single crystal

orientations, strain rate etc. [17]. A considerable research effort has been devoted in the

past decade
mechanical P‘
Them
NiAl. Fastc:
room temper;
emhrittles the
by static str:
Microallofing
NiAl single 0
mechanism fo
testing at amh
Stress and incn
Nil] [23 ~25].
thought to be t

flOm stoichiorr

 

the methanical
Msible to ach
the aboner.

mane“ dUCllI

 

 

may be pOSSlbli
Despite

 

past decade to understand and overcome these shortcomings and to improve the
mechanical properties of NiAl.
Thermal history has been shown to influence the room temperature toughness of
Ni A]. Faster cooling through the temperature region of 673 K to 473 K increases the
room temperature fracture toughness while slower cooling through the same region
embrittles the material. These observations suggest that dislocation mobility is hindered
by static strain aging where interstitial impurities pin the dislocations [18-20].
Microalloying with Fe, Ga or Mo increases the room temperature tensile elongation of
NiAl single crystals [21, 22]. Gettering of interstitial elements is thought to be the
mechanism 'for this improved ductility. Hydrostatic pre-stressing prior to mechanical
testing at ambient temperature and pressure has also been shown to decrease the flow
stress and increase the tensile ductility during subsequent deformation of polycrystalline
NiAl [23 --25]. An increase in mobile dislocation density as a result of pre-stressing is
mought to be the driver for this property enhancement. Crystal orientations, deviations
from stoichiometry, and surface conditions also play very important roles in controlling
the mechanical behavior of NiAl single crystals. Levit et al. [26] have shown that it is
' Possible to achieve tensile elongation as high as 25% at room temperature by optimizing
the above-mentioned parameters and activating a single operative slip system. This
enhanced ductility indicates that sufficient dislocation mobility at ambient temperature
may be possible under specific conditions.
Despite many past studies of dislocations and slip, the exact nature of dislocation
mobility and generation in achieving extensive plasticity in NiAl is not well

characterized. Although the interrelation between increase in ductility and dislocation

mobility is C!
regard has be:
mntmlling the
comprehend t
XiAl.

In ordt
and high pun]
1331315 have b
to induce "301
Were Oriented
room tempcru
ObSmed by C0
motion have
Dislocations g,

dimmed a

 

 

mobility is clearly suggested by many investigators, no systematic investigation in that
regard has been carried out. In order to develop a clear understanding of the mechanisms
controlling the mechanical behavior of BZ NiAl, the present study has been undertaken to
comprehend the role played by dislocation generation and mobility on the ductility of
MA].

In order to understand the role of impurities on dislocation behavior, commercial
and high purity single crystals of NiAl have been used for the investigation. The single
crystals have been cooled at different rates from a homogenization annealing temperature
to induce varying degrees of static strain aging. Following heat-treatment, single crystals
were oriented along different stress axes. Prior plastic deformation was carried out by
room temperature compression. Both deformed and undeformed specimens were
observed by conventional transmission electron microscopy. Dislocation generation and
motion have been studied by in-situ straining transmission electron microscopy.
Dislocations generated in the process of plastic deformation in single crystal NiAl are

characterized as a function of composition, orientation and cooling rate.

 

MAI h
its phssical eh
wide composir
memory efi‘ee:
m the present
behaiior of Ni
all those prop.
properties‘ [Cit
the mechanica
311d eni'ironm.
Particular emp
ductility. Rey
mmdate the
2'1 Phi’Sical Pr

NiAl i,-

”“0 of 1.5.

l673. K) and 5;..-

Kit} is highly

93359833 a mg:

 

2.0 BACKGROUND

NiAl has often been a choice for basic scientific studies in ordered alloys due to
its physical characteristics such as a simple crystal structure, a highly ordered lattice, a
wide compositional range of stability, a variable defect structure and a reversible shape
memory effect. It is important to review the physical and mechanical properties of NiAl,
as the present study is intended to look into the micro-mechanisms of the mechanical
behavior of NiAl single crystals. Since Miracle [6] and Noebe et al. [7] have described
all those properties in great details in recent reviews, only an overview of different
properties, relevant to the dislocation behavior in NiAl, is presented here. A review of
the mechanical properties and elevated temperature properties such as diffusion, creep
and environmental resistance of NiAl follow review of the physical properties. A
particular emphasis is given to dislocations and slip systems and their implications on
ductility. Review of strain aging in NiAl is included as the present study also aims to
investigate the strain aging behavior of high and commercially pure NiAl single crystals.
2.1 Physical Properties

NiAl is an ordered B-phase electron compound with a valence electron to atom

ratio of 1.5. The primitive cubic BZ (CsCl prototype, space group Pm3m) crystal
structure consists of two interpenetrating simple cubic sub-lattices of Ni and Al atoms
(Figure 1). The 82 structure is stable for large deviations from stoichiometry (~20 at% at
1673 K) and significant long-range order is observed up to the melting temperature [27].
MA] is highly ordered and even quenching @ 108 K/s cannot suppress the order. NiAl

possesses a high degree of thermodynamic stability with a negative heat of formation of

J

 

 

 

 

,/
\

 

 

 

0A 0|

Figure 1. 32 structure, also known as the ordered bcc lattice. Two types of atoms (A and
B) occupy the comer positions or body center position respectively.

13.2 kl g-aton
covalent is th
significant bu
direction [6. 7
Bl ;\'
application. h;
the Phase di:
mimlmtly at
materials. Th
and iml’rmed
STOlci’liOmeu-y
Imperature. st
0.2887 nm, Tl
density advam
ds'pendjng 0n t
The eh.
on (lipical or
electron SCanQ.
hill is abOui
allots. SiOlCh

thermal Condhk

 

 

72.2 kJ/g-atom [28]. Bonding in MA] is a mixture of covalent and metallic type, where
covalent is the dominant feature of bonding. High NiAl bond strength results from
significant build up of directional d—bonding charge along the nearest neighbor NiAl
direction [6, 7].

B2 NiAl, a potential candidate material for high temperature structural
application, has some specific advantages over conventional Ni-base superalloys. From
the phase diagram (Figure 2) [29] it is evident that stoichiomentric NiAl melts
congruently at 1911 K, which is at least 300 K more than that for conventional Ni-base
materials. The higher melting point offers advantages of higher operating temperature
and improved creep resistance. The solidus melting point depends strongly on
stoichiometry and drops sharply with deviations from stoichiometry. At room
temperature, stoichiometric NiAl has a density of 5.95 g/cm3, with a lattice constant of
0.2887 nm. The density is at least 3 g/cm3 lesser than the nickel-base super-alloys. This
density advantage could result as much as a 40% weight saving in an aircraft engine,
depending on the rotary motion of the turbine components [1].

The electrical resistivity of stoichiometric NiAl at room temperature is 8-10 p.0-
cm (typical of metals) and increases on both sides of stoichiometry due to increased
electron scattering by constitutional defects [30]. The thermal expansion coefficient of
MA] is about 30% lower than pure Ni, but is very similar to that of nickel-base super-
alloys. Stoichiometry has little influence on coefficient of thermal expansion. The
thermal conductivity of MA] single crystal is 70-80 W/m-K, which is almost eight times

larger than that of conventional nickel base super-alloys and about one-third of aluminum

[1].

l
1800

1600

1400

1200

1000

Temperature (C)

 

 

00362. Phd
[he NIAI alloy

 

 

Wt% Ni

 

 

 

 

0 19 go 30 J so so 7p 89 9.0 100
1800 ' ' F T 1' 1 r I i I
1638°C
1600 l- a
'- \ 1455°c
8 1400- ' ; -/' a
.. MM
2 1200b Ni -
3 1 33°C _]
s l .
a 1000-
5
’- d

o
I].

'4‘ i ' .-
2" AlaNis «1 b-AlNi

1 I [J 3 i

40 50 60 70 i 0 90 100

At% Ni Ni

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2. Phase diagram illustrating the phase stability as a function of temperature for
the Ni-Al alloy system [29].

The n
Thus. when “I
density is ex;
takes place 11"
Al and Ni ato:
X10311 [o].
and the stoicl:
and gray for
the sharp chart

NiAl 1?

0n orientation

lhls i5 measur

eonsram ‘A‘
than StrOngly ;

‘ v

I‘3'"1156rature (\

 

The Yc

B a function 0.

time E and G

the X’Oungvs m‘

01qu1.) on

S’r . .

The nickel atom is heavier and smaller in size in comparison to the Al atom.
Thus, when Ni is substituted for A1, a reduction in lattice parameter and increase in
density is expected. On the contrary, a decrease in both lattice parameter and density
takes place for Al-rich alloys (Figure 3) [7]. The maximum interaction energy between
A1 and Ni atoms is observed at the stoichiometric composition and is of the order of -11.5
X 10'20 J [6]. The visible appearance of NiAl is heavily dependent on the alloy chemistry
and the stoichiometric bright blue color changes to yellow or bronze for Ni-rich alloys
and gray for Al-rich alloys; interband transitions associated with Fermi surface results in
the sharp change in color [7].

NiAl is crystallographically anisotropic and the elastic properties depend strongly
on orientation. Stoichiometric NiAl possesses an elastic anisotropy factor (A) of ~3.3;

20,,

this is measured by Zener's parameter, A = , where C is the elastic stiffness

ll — 12

constants. ‘A’ for NiAl is comparable with Cu (~3.21) and Ag (~3.01) and much lower
than strongly anisotropic material such as B-brass (~8.49). ‘A’ is weakly dependent on
, temperature (~2.94 at 873 K) and strongly dependent on composition (A = 2.56 for Ni-
52% Al and A = 5.67 for Ni-45% A1) [31].

The Young's modulus (E) and shear modulus (G) of NiAl have been determined
as a function of temperature [7]:

E = 199.8 - 0.04 (T) and G = 76.6 - 0.0.17(T),

where E and G is in GPa and T is in K. The single crystal orientation strongly influences
the Young's modulus, with values of 96 GPa, 187 GPa and 275 GPa for <011>, <100>,
and <111> orientations respectively for nearly stoichiometric NiAl [31]. The yield

strength is also strongly dependent on alloy chemistry and shows a minimum at the

AECV 5559mm 005m...

Variation of Lattice Parameter and Density with

 

 

Stoichiometery
0.289 7
0.2885 «1
6
A 0.288
r E A
7: 0.2875 2’
l 9 , 5 5
l g 0.2871 g
, e 3’
1 330.2865« 4;»
(I)
, .3 0.286 ., 5
1 fi 0
—J 0.2855r 3
0.285 «
0.2845 2

 

 

 

43 45 47 49 51 53 55 57 59 61 63
Nickel Concentration (at%)

Figure 3. Lattice parameter and density of NiAl vary as a function of stoichiometry [7].

stoichiometric

mdcpendClli c

Ni-Al is fou
susceptibility
weakly relatet
2.3 Mechanic.

All oi~
strongly by th

(point defects

 

 

the focus of
crystalline def
ductility. fract
2.2.1 Crystal 1
Depen
line, and sur-
meChemical pi
2.2.1.1 Point
NiAi
have mple dc
{35}: 0 and l
defect free la

defects, m0

stoichiometric composition [32-34]. Poisson's ratio (v) have been found to be almost
independent of temperature, T:
v = 0.307 + 2.15x10'5 T [6].

Ni-Al is found to be non-magnetic or Pauli paramagnetic, with a low magnetic
susceptibility of ~2X10'2 to 6X10“2 emu/gm; magnetic susceptibility of NiAl is very
weakly related with temperature and composition [7].
2.2 Mechanical Properties

All of the mechanical properties of NiAl are structure-sensitive and influenced
strongly by the defect structure, whether it is localized to the proximity of a few atoms
(point defects) or extends through the microscopic regions (lattice defects). Considering
the focus of the present study on the mechanical behavior of NiAl, a review of the
crystalline defects in NiAl will be presented first, which will be followed by slip systems,
ductility, fracture, and strain-aging.
2.2.1 Crystal Defects

Depending on the scale of presence, the crystalline defects are classified as point,
line, and surface or planar defects. These three defects and their influence on the
mechanical properties of MA] will be reviewed in the following sub-sections.
2.2.1.1 Point Defects

NiAl exhibits both constitutional defects and vacancies. MA] is considered to
have triple defect structure i.e. when A atom occupy the [3 site, the or site remain vacant
[35]; or and [3 are the sites normally occupied by the A and B atoms respectively in a
defect free lattice. The term triple defect refers to the simultaneous presence of three

defects, two vacancies on or site and one A atom on [3 site at the stoichiometric

10

wmpOSlliOil.
stoichiometr}
Al atoms on ‘
are formed in
and strong tt
constitutional
minimize the
vacancy form
Simmetn'c wit

Beside

thermal vacar

 

stoichiometei '
mpemrures,

dam).- 0f 00
energy and sli
the'mial Vacar
modulus [40]
Produced by

[~11], “mm;
rapid qUenchl‘

“nearing at 1

composition. Usually the Al sites are fully occupied. With deviations from
stoichiometry Al atoms never reside on Ni sublattice. Any excess Ni atom substitutes on
Al atoms on the Al sublattice on the Ni rich side. Vacant sites or constitutional defects
are formed in the Ni sublattice with excess Al [7]. Because of the high ordering energy
and strong tendency to avoid the same defect species in the nearest neighbor site,
constitutional defects in NiAl exhibit somewhat ordered structures that also help to
minimize the strain energy associated with the introduction of a defect [36]. The ease of
vacancy formation differs in the two sub-lattices and as a result vacancy formation is not
symmetric with stoichiometry [37].

Besides constitutional vacancies it is possible to quench a high concentration of
thermal vacancies that leads to strengthening in NiAl [38]. Near the melting point of
stoichiometeic alloy, 2% of thermal defects may exist. After quenching from moderate
temperatures, 0.5 - 1% of such vacancies can be retained [39]. The presence of a high
density of quenched-in vacancies might be a consequence of relatively low formation
energy and slightly higher migration energy for vacancies in NiAl. These supersaturated
thermal vacancies increase the yield strength without significantly affecting the elastic
modulus [40]. Vacancies offer a much greater resistance to dislocation motion than that
produced by substitutional atoms, and the increase in hardness can be as high as 20%
[41]. Thermal vacancies are susceptible to agglomeration and form faceted voids after
rapid quenching from the melt [3 8, 42]. Dislocation loops and helices are formed during

annealing at lower temperature by thermally induced vacancies [33, 41-43].

11

2.2.1.2 Line I

The st
ofan)‘ materi
in strong nea:
order of the 1
Kit]. All 01
vectors occur
into partials.
Structure taker
Energy and th
screw orient;
configuration
extra energy i
SCI”00911131

NOH‘C‘t
decompositior
p”170$ch that

ion“ an a<ti

 

a<iio>{]10,

e‘) .1"
d ‘0‘10 dCCrea

ac .
Summer-“1y be

A 8m:

[ll] “10; Slip r

 

2.2.1.2 Line Defects

The study of dislocations is important in understanding the mechanical behavior
of any material. The large ordering energy in MA] does not favor slip vectors that result
in wrong nearest neighbor atoms. Three basic translation vectors that do not disrupt the
order of the lattice are a<011>, a <100> and a <111> where a is the lattice constant of
NiAl. All of these dislocations have been observed in NiAl, although non-<001> slip
vectors occur under specific circumstances [6, 7]. The dislocation cores partly dissociate
into partials, and with increase in planar fault energy a transition to undissociated
structure takes place [44]. Dislocations with <001> Burgers vectors have the lowest line
energy and the edge orientation of <001> dislocation has slightly lower energy than the
screw orientation (Figure 4) [41]. a<001> dislocations within 25° of the screw
configuration are elastically unstable on {110} and {001} planes [6, 45]. Thus, some
extra energy is required to force an elastically stable a<001> mixed dislocations into the
screw orientation.

Non-edge a<110> dislocations on {110} planes have sufficiently low energythat
decomposition into two a<001> dislocations is not expected [6]. However, it has been
proposed that an a<110> edge dislocation lying on an {110} plane may decompose to
form an a<001> edge dislocation and an a<001> screw dislocation. [40]. An
a<110>{110} edge dislocation can dissociate into two a/2<111> dislocations resulting in
a 20% decrease in energy and require the antiphase boundary (APB) energy to be
sufficiently low [6, 46].

A small amount of elastic instability is associated with a<111> edge dislocations

on {110} slip planes but not on {1 12} planes [45, 47]. These a<1 l 1> dislocations may

12

Llne energy

Line energy

 

[001] [111] [110] [112] [001]

(.0

 

  

edge

   

(1 in) Plane

N
01

[111]

 
 

 

 

 

 

 

 

 

1 5 2*
i g 15 —— screw [110]
1 g ‘——_ *-"""
edge
1 0.5 r [001] '
o
45 135 180
(9) Degrees
10011 11011 11001 [10T] [00T].
37—~~—~ 222 __2 -22 ,223 222 ~ _._2. '~—--
1
2.5 7
1 >. ; (010) Plane edge .
9 2 ‘1 ,
i g 1.5 . screvv [101] ,
l g 1 1
' 0'5“? edge [0011 ‘ ' i
0 ..2, 2 . 7... 2,42,. 2
0 45 90 135 180
(0) Degrees

Figure 4. Dislocation line energy in stoichiometric NiAl at room temperature.
Dislocation line directions are measured from [001] direction [41].

dissociate 11‘
appropriate 1:
loops are oh
character and
2.2.1.3 Planar

In gen
to ditierentia
energies exist
onstals, it is r
faults likely It

a1103153 - s.

 

by clustering
i001} Planes [

“W [45. 4'

dissociate into a<001> + a<110> dislocations or three a<001> dislocations of
appropriate line direction on the {110} planes [6]. Several types of sessile dislocation
loops are observed after room temperature deformation [7]; these loops are edge in
character and lie in {001 } planes with a<001> Burgers vectors [7, 43, 48].
2.2.1.3 Planar Defects

In general, stoichiometric NiAl does not exhibit any planar defects. The inability
to differentiate splitting in superlattice dislocations strongly suggests that high fault
energies exist in NiAl. Based on theoretical analyses of possible planar faults in B2
crystals, it is revealed that a/2<111>APBs on {110} and {112} planes are the only stable
faults likelyto occur in NiAl [49]. Deviation from stoichiometry in slow-cooled Al rich
alloys (53 - 54 at%) gives rise to planar faults. These planar faults are suggested to form
by clustering of point defects, which are either vacant Ni sites or excess Al atoms on
{001} planes [50]. In general, deformation induced planar faults have not been observed
in MA] [45, 47].
2.2.2 Slip Behavior

The deformation and slip behavior of NiAl single crystals is strongly dependent
on the orientation of the single crystal. The literatures generally classify orientation as
hard and soft. Orientations close to <001> are considered 'hard' because <001> slip
vectors have zero or near zero resolved shear stress and the stress necessary for
deformation is several times greater than for any other orientation to operate alternate slip
systems [7]. 'Soft' orientations include all non<001> directions of loading, where <001>

slip vector dominates.

14

NiAl
orientations-
dislocation g
the primary s
[32, 45, 50].
the operatior
orthogonal ;
slip on {110
dislocations

\l'nilewski e1

 

deformed beti
than for <00};

Defon-
and {“0} pie
M47, 55, 5(
<110> and <C
(‘77 K), 1h 6 (
At higher ten
deformation 1

1011101113“ 011 C

NiAl exhibits two significantly different types of slip behavior [50-63]. For 'soft'
orientations <001> slip vectors dominate and do not depend on temperature. Considering
dislocation glide in non<001> oriented single crystals, <001>{110} and <001>{001} are
the primary slip systems and have been confirmed by various TEM and slip trace studies
[32, 45, 50]. In the temperature range of 300 — 1273 K, Ball and Smallman [52] found
the operation of <001>{110} slip in soft orientations and observed cross-slip on
orthogonal {110} planes. Although the critical resolved shear stress values for <001>
slip on {110} or {001} are similar, relative mobility calculations revealed that <001>
dislocations preferentially glide on {110} planes to {001} planes [52, 59, 60].
Wasilewski et a1. [32] observed cube slip <010>{100} in <110> oriented single crystals
deformed between 473 and 973 K; <001>{001} slip occurred at slightly higher stresses
than for <001>{110} slip [52, 59].

Deformation of hard oriented single crystals take place by <111> slip on {112}
and {110} planes between 77 K and 300 K [45] and between room temperature and 600
K [47, 55, 56]. Depending on the deformation temperature, a combination of <111>,
<110> and <001> slip vector dominates for hard orientations. At very low temperatures
(~77 K), the observed slip direction is <1ll> on {110}, {112} or {123} planes [45, 53].
At higher temperatures (>600 K), <111> dislocations become thermally unstable and
deformation has been ‘ reported to be the result of motion of <001>, <110>, or a
combination of glide and climb of <001> and <110> dislocations [51, 54-56, 61].

Sample geometry plays an important role in determining deformation behavior of
hard oriented single crystals. Both kinking and uniform deformation takes place by

<111> slip on {112} and {110} planes during deformation of hard oriented single

15

crystals [7].
oonfinned as
2.4 and 3) in
56]. This is
ratio of the c
promoted the
ratio [63].

Defor
dislocations .

independent i

 

 

According to.
{We independ

Ill \lAi, n0 ‘

crystals [7]. However, kinking has never been reported in any tension test and is
confirmed as an artifact of an unstable testing geometry (length to diameter ratio between
2.4 and 3) in compression tests, which proceeds by movement of <001> dislocations [54,
56]. This was supported by the presence of only <111> dislocations when the aspect
ratio of the compression specimen was kept ~2 [7, 45]. Also, alloying additions of Cr
promoted the activation of <111> slip over deformation by kinking for a constant aspect
ratio [63].

Deformation of polycryStalline NiAl generally occurs by the operation of <001>
dislocations as in the case of soft oriented single crystals [57, 58]. Only three
independent slip systems are available as a result of this predominantly <001> slip.
According to Von Mises criterion for plastic deformation of polycrystalline materials,
five independent slip systems are required to accommodate any arbitrary shape change.
In NiAl, no extra independent slip system is provided by cross-slip [52]. <111> slip
could give rise to five independent slip systems, which is sufficient to produce an
arbitrary strain and extensive crack free deformation without change in volume.
Although <111> slip satisfies the requirements for generalized polycrystalline plasticity,
no tensile ductility has been reported in Mn or Cr containing polycrystalline NiAl at
ambient temperature; a change in fracture mode from intergranular to transgranular
cleavage has been observed with <111> slip [7]. Attempts to alter the slip vector of NiAl
by alloying addition had not been successful as majority of the investigations reveal
almost exclusively <001> dislocations. In extruded alloys <110> dislocations are
observed in {011} planes, which are considered as a product of dislocation reactions and

have very little potential to increase plasticity [7].

l6

2.2.3. Yield 8

lt has
strongly on 1
room temper
extended on
stress decreas
single crysta
orientations [
polycrystals L
hard oriented
and 300 K 81
hard and sot
dlfiusion start

Like (t
(Figure 5) [7
reaches the n
Sigm‘ficantly }
hemsen 300 .
MA] inCTfi’LsQ-r
ofdifietem te:
[1 5,7, 66], I

58321 Egan 011 at

‘0-036 nm in ‘

 

2.2.3 Yield Strength

It has been established that the yield stress (0,) of polycrystalline NiAl depends
strongly on the temperature of deformation below room temperature but at or close to
room temperature an atherrnal regime is observed [64]. These atherrnal plateaus are
extended over 300 K, and generally occur between room temperature and 800 K. Yield
stress decreases slowly with increasing temperature beyond this plateau temperature. For
single crystals, the yield stress is much higher in hard orientation, than any other
orientations [7]. Soft oriented single crystals exhibit similar temperature dependence like
polycrystals and the yield stress decreases continuously with increasing temperature. For
hard oriented single crystals, yield stress is a strong function of temperature between 77
and 300 K and deformation takes place by <110> and <001> slip; above 900 K, both
hard and soft oriented single crystals and polycrystals behave similarly as the bulk
diffusion starts to dominate [7].

Like other properties of NiAl, the yield strength depends heavily on stoichiometry
(Figure 5) [7]. The flow stress and hardness reaches the minimum and the ductility
reaches the maximum value at stoichiometric composition. The flow stress increases
significantly by deviating from stoichimetry. The yield stress is moderately constant
between 300 and 600 K and decreases sharply above 600 K [64, 65]. The flow stress of
NiAl increases highly in the presence of solutes. Table I and figure 6 describe the effects
of different ternary additions on solid solution strengthening of MA] at room temperature
[7, 57, 66]. B and Zr give a much higher hardening rate because of grain boundary
segregation and second phase particle formations. The calculated interstitial site radius is

~0.036 nm in MA] and it is predicted that smaller atoms like B and C are more probable

l7

30.): 305mm 0.0.;

Figure 5. t

100 Yield stress changes with stoichiometry
1

1000 . . 1 0 600K
900 1 I 873 K
300 K
80° x 673K ‘
700 1:: 973K ‘
600

500

 

    
 

20 MPa/at% x

Yield Stress (MPa)

300
200
1 00

 

 

 

46 48 50 52 54 56 58 6o 62 64
Nickel (at%)

Figure 5. Yield stress of NiAl varies strongly as a function of stoichiometry [7].

18

TABLE I. Solid Solution Strengthening in NiAl [7]

 

 

 

 

 

 

 

 

Element Concentration Atomic Radii Ac Goldschmidt
" (MPa/at%)
AC ..
(nm) radii, nm

Boron 0.04 0.1 17 4500 0.093
Zirconium 0.05 0.216 4000 0.157
Carbon 0.1 1 0.0091 >1700 0.084
Chromium 1.00 0.185 258 0.125
Beryllium 0.24 0.140 100 0.1 1 1

 

 

 

 

 

19

 

 

RIC
t E QC:

F2 CECQtLrI

c

S:

m \ run.

E...

Figure 6. R

ale crystz

Sin;

Hardening rate changes with size of the solute

 

6000

5000

4000

3000

2000

1 000

Change in Hardening rate (MPa / at%)

 

 

 

 

 

0 0.05 0.1 0.15 0.2 0.25 0.3
Goldshmidt radii (nm)

Figure 6. Relationship between hardening rate and size of the solute elements in NiAl

single crystals [7].

20

to occupy I?

 

stoichiometr'
Ni or Al-n't
vacancies in
considered it

The _\
grain size co
Theyield str
COHVCTde g

alloys. Ho

 

increasing ta

For p
or below 50
dialed at ele
mill rate 5e]
T< 1350 K);
PU SeCOnd to
3.3.4 Ductilir

The n

omen-31 is it

to occupy the interstitial sites [66]. Discontinuous yielding has been reported in off-
stoichiometric alloys. Serrated stress strain curves or presence of yield points for either
Ni or Al-rich alloys are considered as discontinuous yielding; while the presence of
vacancies in Al-rich alloys are responsible for serrations, dynamic recrystallization is
considered to be the driving mechanism in Ni-rich alloys [67, 68].

The yield stress is affected by the grain size in a complex manner. Decreasing the
grain size can decrease or increase the yield stress or it may not be impacted at all [7].
The yield stress is almost independent of grain size for nearly stoichiometric NiAl alloys;
conversely grain-size influences the yield stress significantly for non-stoichiometric
alloys. However, this influence of deviation from stoichiometry decreases with
increasing temperature and becomes negligible above 800 K [7].

For polycrystalline NiAl alloys, yield stress is marginally affected by strain rate at
or below 500 K, moderately influenced at intermediate temperatures, and significantly
altered at elevated temperatures [57]. Rozner and Wasilewski [69] observed the largest
strain rate sensitivity of cube oriented single crystals at intermediate temperatures (750 <
T < 1350 K); the increase in flow stress was ~ 30% for a change in strain rate from 104
per second to 10'3 per second.

2.2.4 Ductility

The most serious limitation associated with the application of MA] as a structural
material is its lack of room temperature ductility. Until the work of Levit et al. [26]
limited (< 2.5%) tensile ductility was reported for soft oriented single crystals below 500
K. Sofi oriented NiAl single crystals undergo a brittle to ductile transition at

temperatures ~473 K (about 0.25Tm) [1]. Different factors such as insufficient

21

dislocation

 

slip contrilf
single cryst;
As a result 1
crystals ma}.
Roor
stoichiometr

less than 1"

 

)ield strengtl|
MlCri
Crystals sign
be achieved
it; Ga or M
in ductility
film in lnl
interstitial 91
[his “men
understood_
Hard
16mm. ature
9031.313 be] 0
of a<1 l l> c'

dislocation sources, lack of dislocation mobility, low fracture stress, and heterogeneous
slip contribute to this limited ductility at relatively low temperature [11]. As-grown
single crystals contain mostly edge dislocations, which typically lie on cube planes [40].
As a result of low concentration of mobile dislocations on {110} slip planes, NiAl single
crystals may be dislocation source limited [40, 70].

Room temperature plasticity of NiAl is heavily dependent on the alloy
stoichiometry. The ductility is maximum at stoichiometric composition and deviations of
less than 1% from stoichiometry results in brittleness and a corresponding increase in
yield strength [7].

Micro-alloying additions have been shown to improve the ductility of NiAl single
crystals significantly. Darolia et a1. [1, 21, 22] observed that a tensile strain of 6% could
be achieved for soft oriented single crystals by doping with approximately 1000 ppm of
Fe; Ga or Mo exhibit similar effect but not as pronounced as Fe (Figure 7). This increase
in ductility shows that alloying with more than 0.5% of these elements has no further
effect in increasing the ductility and in fact the benefit might be lost. Gettering of
interstitial elements or slip homogenization is thought to be the governing mechanism for
this tremendous increase in ductility, although the exact mechanism is yet to be
understood.

Hard oriented single crystals are brittle and often fail without yielding at room
temperature [55]; limited tensile ductility is observed for hard oriented NiAl single
crystals below 600 K [17]. The little observed uniform deformation occurs by the glide
of a<111> dislocations [45, 55]. The ductile brittle transition for hard oriented single

crystals occurs at ~600 K [7]. Above that temperature, glide of a <110> dislocations

22

...c 4 n5 n _
1r. Fist—am OmumG—nw &\O

7 ill l ,- l

 

 

 

Microalloying additions improve the ductility of NW single crystals

 

   
 
 

 

 

 

 

 

 

 

 

 

6.
l
5 0M0
E 4 DGa
m A Fe
9* l
°\° 2 HA? ,
1 F‘L_ 42‘: 4h
- l l l
0 as 1 1.5 ' 2

At % Ternary Element Addition

 

 

 

Figure 7. Room temperature ductility of <110> oriented NiAl single crystal with
microalloying additions [1].

23

takes place.

.1

strain) [53,.

 

sufficient tr
predominan
rotation [6. .

Sim:
ductility at r
limited by 1'

ll]. Hydros

 

all'llOSpheric
increase in

attributed to
inclusions or

2.2.

LI-

Fractur

Fract

h is lmPOrta

1. 1 -
....sr pram

takes place, which leaves a<001> debris after a small amount of deformation (~l.5%
strain) [55]. Further deformation results in a rotation of the deformation axis that is
sufficient to activate a<001> glide dislocations; glide of a<001> dislocations become
predominant afler no more than ~6% strain, due to deformation induced crystal axis
rotation [6, 55].

Similar to soft oriented single crystals, polycrystalline NiAl exhibits limited
ductility at room temperature [65, 66, 69, 71]. The ductility of polycrystalline MA] is
limited by the incompatibility of shape changes of neighboring grains [55, 65, 66, 69,
71]. Hydrostatic pressurization between 300 and 1400 MPa prior to mechanical testing at
atmospheric pressure (0.1 MPa) decreases the flow stress by approximately 40% with an
'increase in tensile ductility [23-25]. This change in mechanical behavior has been
attributed to increased dislocation generation at areas of local shear stress such as
inclusions or other material hetero geneities.

2.2.5 Fracture Behavior

Fracture behavior of any material is a critical index for its toughness and ductility.
It is important to review the fracture mode and fracture toughness of MA] to develop a
complete understanding of the mechanical properties.
2.2.5.1 Fracture Mode

In general, room temperature fracture in NiAl occurs without any stable crack
growth. Stoichiometric NiAl single crystals typically fail by cleavage on {110} planes
irrespective of orientation [11, 64, 66, 71, 72]. This cleavage behavior is a consequence
of lower (30%) cleavage energy on {110} planes than on {001} planes [73].

Polycrystalline materials primarily fail by intergranular fracture at low temperatures [65,

24

 

or, 74}. r

 

imperaturc
grain size.

alloys or al.

NiAl

 

 

 

extensive m'
home too.
room tempc
shows limi
IOUghnegs r
tougliness \
The
pol-lm'Stal
0115111311011
mpfiramr
“1’3le “i
1011th egg
.fier 0f \1
The
Maren als a

”Idem in

66, 74]. Transgranular cleavage takes place at and above 673 K [65]. At ambient
temperatures, the tendency towards transgranular fracture increases with increasing the
grain size, strain rate and deviations from stoichiometry [6, 75]. Non-stoichiometric
alloys or alloys with ternary or quaternary additions, which are not known to exhibit
room temperature ductility, fail in a predominantly transgranular mode [75].

2.2.5.2 Fracture Toughness

NiAl displays extreme brittleness at ambient temperatures. In absence of
extensive room temperature tensile elongation, ductility of MA] is often measured by
fracture toughness tests, which also represent the internal resistance to cracking. The
room temperature fracture toughness of polycrystalline NiAl lies between 4 - 6 MPa and
shows limited dependence on grain size and stoichiometry [76-78]. This fracture
toughness value is similar to many ceramic materials such as A1203 with fracture
toughness values in the range of 5 - 6 MPan [7].

The fracture toughness values of MA] single crystals are similar to single phase
polycrystalline material and have been observed to vary significantly with crystal
orientation and temperature. The plane strain fi'acture toughness (ch) at room
temperature has been determined to be 4.5 MPa/m, 5.1 MPan and 8 MPan for single
crystals with <110>, <111>, and <001> stress axes respectively [6, 78, 79]. Fracture
toughness generally increases with increasing temperature and above 500 K toughness
value of ~1 5 MPan has been reported for soft oriented single crystals [6].

The ch increases between 573 and 673 K to about 10 MPan for as-cast

materials and to about 50 MPan for zone-refined materials [76]. This signifies an

increase in fracture toughness with alloy purity at elevated temperature, although room

25

 
   
   
   

[mpClfllUl'L
material.
2.2.6 Strain

NiA:
that the fro
controlled
elongation
immune
crystals are
disappears v
crystals at 6
in substanri;
by slow CQC

Simprible

temperature fracture toughness for the zone-melted material remains as low as original
material.
2.2.6 Strain Aging

NiAl is has a high susceptibility to strain aging. Hack et al. [18 - 20] observed
that the fiacture resistance of NiAl single crystals could be significantly improved by
controlled heat treatments. Furnace-cooled NiAl single crystals exhibit a tensile
elongation of 1% and a fracture toughness of 2.4 MPan at room temperature. The room
temperature tensile ductility and KK: increase to 7% and 17 MPan respectively if the
crystals are re-heated to 673 K and air-cooled. The beneficial effect of fast cooling
disappears when the crystals are firrnace-cooled from 673 K. Creep bending of single
crystals at 673 K and subsequent annealing at 473 K followed by fumace cooling results
in substantial toughness [18]. These observations indicate that static strain aging occurs
by slow cooling, whereas dislocations generated by high temperature deformation are not
susceptible to strain aging.

NiAl exhibits severe strain aging as interstitial atoms tend to segregate to the core
of the mobile dislocations. Many of the conspicuous features of strain aging phenomena
such as occurrence of yield points, discontinuous yielding, low strain rate sensitivities
and plateaus in the temperature dependence of yield stress, and work-hardening rates are
present in the deformation of single crystal and polycrystalline MA] [80]. Weaver et al.
[81] observed serrated flow in the presence of C (>150 ppm) and Si (>1500 ppm).
Silicon enhances the serrations in the stress-strain plots by increasing the activity as well
as diffusivity of carbon in NiAl. Brzeski et al. [20] observed serrated yielding during

compression testing between 373 K and 473 K, in contrast to smooth yielding at room

26

[gmpCfalUlt
between 70
and 0 are t
typical of it
dtnamic str,
is the main
help to remr
Yield point:
hydrostatic 1
annealing of
3.3 Dillusioi

In at

 

 

Considered a
CFCle consist
mechanism [
activation 3n
The dl 1Ll‘lJSlOr
lilCreaSe in

concentratior

Et’aduany der

temperature. In a separate study, Brzeski et al. [82] have also observed serrated flow
between 703 K and 873 K. Calculated activation energies for the process suggests that C
and O are the probable agents for this serrated yielding behavior. These serrations are -
typical of interactions between dislocations and interstitial solutes, commonly known as
dynamic strain aging. Weaver et al. [83] have shown that carbon, not oxygen or nitrogen,
is the main interstitial element responsible for strain aging. Alloying with Ti and Mo
help to remove the yield point totally for their strong affinity towards carbon [81, 84, 85].
Yield points could be suppressed by faster cooling from higher temperatures, and by
hydrostatic pre-stressing [83, 86]. Recovery of yield points is observed by subsequent
annealing of pre-stressed material at 700 K [87].
2.3 Diffiision

In absence of divacancy mechanism, next nearest-neighbor vacancy jumps is
considered as the important diffusion mechanisms in NiAl; in addition to that a six jump
cycle consisting only of nearest-neighbor vacancy jumps is also proposed as an operative
mechanism [88, 89]. A minimum in the diffusion co-efficient and maximum in the
activation energy is observed near stoichiometric composition (Figure 8 a and 8 b) [7].
The diffusion coefficient increases significantly with deviations from stoichiometry. This
increase in diffusivity with increase in Al is caused by the presence of a high
concentration of constitutional vacancies and thermal vacancies. The activation energy
gradually decreases to the Ni-rich side and drops sharply for Al-rich compositions [89,
90]. The rapid decrease in activation energy is a result of faster diffusion of Al in Al-rich
NiAl and driven by the energy of motion of constitutional vacancies. The higher

activation energy on the Ni-rich side consists of both the energies to create and to move

27

 

:oEEJ: >Uuecm Cetu>to<

 

 

 

 

 

 

 

 

 

 

 

 

10'7 .. .
8
fl 0 .0
x -8 d
‘E 10 r a 60 Co
3 O C] .0 D
g 10'9 i" .0 Cl ..8 A O114In
g 10 .0838 0&3 O A Olnter-
g 1 0' + QO‘QAAD diffusion
o V63m
10’”
40 50 60 70
Ni (at%)
(a)
400
7 T’ .
. :5 350 lnterdrffusron
' E |I63 NI
l i 300 ' 600°
i ; ,, .1114 "t-____
‘ c o . . .
1 I: . °
1 .2 200 0 . I . O .
3g? I o 9 X
.. 15o .
‘6 o
. <
A 100
45 50 55 60 65 70
Ni (at%)
(b)

Figure 8. Diffusion coefficients (a) and activation energy (b) for diffusion of various
species in NiAl as a function of stoichiometry [7]

28

vacancies.

and 265 it.

 

activation L
[140 hi 'mo
suggest tha
[90], which
2.4 Creep
The

Creep rate is

 

Creep-rate (l

actuation er

Where K is
antiphase b
Itimperature
Sugg't‘Sled 1
am.VfillOn C
activation er

The
stoicthmen

b." it ShOrt [-

vacancies. Figure 8 b shows that the average activation energy lies in the range of 185
and 265 kJ/mol and does not vary widely with composition. These observations of
activation energy do not conform to the independent energy calculation for formation
(140 kJ/mol) and migration of vacancies (~160 kJ/mol) [7]. The sum of these two values
suggest that the activation energy for diffusion in NiAl should be closer to 300 kJ/mol
[90], which implies short-circuit diffusion that was not captured in the earlier models.
2.4 Creep

The creep properties of MA] are not strongly dependent on stoichiometry and the

creep rate is uniform in the composition range of 45 and 52% Al [7]. The steady state

creep-rate (a) is expressed [6] as a function of applied stress (0), stress exponent (m) and

activation energy (AQC) of creep:

(e = Ka’"[exp(—AQC /RT)]
where K is a constant associated with microstructural effects and stacking fault or
antiphase boundary energies, R is the universal gas constant and T is the absolute
temperature. The activation energy for the creep of near stoichiometric NiAl has been
suggested to be a function of stress, temperature and composition and the average
activation energy for creep in NiAl is ~314 kJ/mol [7]. This is in close agreement with
activation energy measured in some diffusion experiments [89, 90].

The steady state or secondary creep region is observed between 1000 to 1300 K in
stoichiometric NiAl single crystals and non-stoichiometric polycrystals that is preceded
by a short period of normal primary creep [6, 51]. Information about tertiary creep in
NiAl is not conclusive. Two-phase alloys of NiAl are reported to be stronger than the

single phase NiAl alloys in creep and stress rupture [1, 91].

29

Slit

much infer

 

 

carbon (Ml
creep strer‘.
precipitatio
decrease thu
Mici
commonly

det'elOped n

(row?) a

 

{001} and {
Perlcil glide
“(ODD disl
deformation
dislocations
iron It “h
and1300 K;
2.5 Em'irom

Ninr
based high 1
A1303 SCale
rcsiStanCe of

'ch Steady Sl

Stress-rupture values of binary NiAl along an <110> axis and of polycrystals are
much inferior to those of Ni-based superalloys [6]. External factors, like addition of
carbon (~420 ppm), significantly decrease the activation energy of creep and thus the
creep strength [92]. The creep rate above ~1000 K is significantly affected by
precipitation of vacancy/impurity complexes [93]; substitutional additions like iron
decrease the steady state strain rate [94].

Microstructural features such as dislocation tangles, dipoles, and loops are
commonly observed in creep-deformed materials [51, 95, 96]. Above 1200 K well
developed networks and subgrain formation takes place and the dislocation density is low
(10'7cm'2) within the subgrains [94, 95]. Extensive cross-slip is evidenced between
{001} and {110} slip planes [95, 96]. Deformation in ‘soft’ single crystals takes place by
pencil glide of a<001> screw dislocations between 748 K and 1048 K; presence of
a<001> dislocation loops are supposed to be a result of double cross-slip [6, 96]. Creep
deformation in hard orientated single crystals occurs by the motion of a<110>
dislocations [51]. Diffusion of Al plays a major role in the climb of dislocations below
1100 K, while deformation of NiAl is controlled by dislocation climb between 1100 K
and 1300 K; viscous glide or solute drag becomes operative above 1500 K [6].

2.5 Environmental Resistance

NiAl is highly resistant to oxidation and this property is utilized to develop NiAl-
based high temperature coatings for Ni-base superalloys [7]. Formation of protective
A1203 scale and growth of the oxide films at high temperature gives rise to oxidation
resistance of NiAl. There is a transient growth stage during build-up of the scale before

the steady state region sets in under isothermal conditions. Transition oxides like NiO

30

and NiAlgt
rate of Al:(
and the up
oxidation i~|
when the al

NrA
difference
spallation ['

and thus tl‘.

 

 

 

 

 

 

 

conditions.
Pifliously r
ill Rare e;

alloys [99].

and NiA1204 form during the transient stage and at lower temperatures when the growth
rate of A1203 is slow [97]. The growth rate of the A1203 scales follows a parabolic curve
and the type of the scale (0-A1203 or a-A1203) depends on the temperature [7]. Rate of
oxidation is a function of alloy chemistry and the rate reduces by an order of magnitude
when the aluminum concentration is increased from 42 to 50 wt% [98].

' NiAl performs satisfactorily for the cooling portion of cyclic oxidation when
difference in thermal contraction coefficients of metals and oxides results to scale
spallation [7]. The spalling induced damage increases with increasing number of cycles
and thus the lifetime is shorter during cyclic oxidation in comparison to isothermal
conditions. Continued cyclic oxidation results in Al depletion and as a consequence
previously mentioned transition oxides (NiO and NiA1204) form leading to shorter lives
[7]. Rare earth elements are reported to improve the oxidation resistance of NiAl based

alloys [99].

SUMMARY

This background information suggests that the limitation in plastic deformation in
NiAl at room temperature is not necessarily a consequence of a lack of dislocations, but
rather a lack of mobile dislocations. Dislocations may be pinned by interstitial atoms in
strain-aged single crystals, resulting in insufficient mobile dislocations for plastic
deformation. Observation of extensive elongation in high purity single crystals (lower
concentrations of impurities than in commercially/conventionally pure materials) further

reinforces this concept. Similarly, improved ductility in micro-alloyed single crystals

31

may teller

 

decrease in

Stui
are the mil
found to 51
these two 5
on {NO} p
energy cons
satisfy Von

materials. 1

 

resulted ini
mmhg'
Des}
ductility en}
this relation
mm” lens
Separaret). c
found that
dislocation
i“l-V’Ctt'sraii
mobile in a
chimed th

WWsm

may reflect gettering of interstitials. In pre-stressed polycrystalline materials, the
decrease in flow stress results directly from an increase in mobile dislocation density.

Studies of dislocations and slip systems in MA] suggest that <100> dislocations
are the most commonly observed dislocations. These <100> dislocations have been
found to slip on {110} and {001} planes [6,7]. The critical resolved shear stresses of
these two slip systems are also similar. Sometimes <110> dislocations are also observed
on {110} planes when dissociation into two <001> dislocations is not expected due to
energy considerations. <111> slip has received considerable attention for its capability to
satisfy Von Mises criterion requiring five independent slip systems in polycrystalline
materials. But irrespective of the method of generation, <111> dislocations have never
resulted in increased plasticity. The critical resolved shear stress for <111> slip may also
be too high to cause yielding prior to fracture.

Despite the fact that the interrelationship between dislocation mobility and
ductility enhancement is clearly suggested in some studies, no systematic investigation on
this relationship has been carried out. In order to study the dislocation mobility in MA],
in-situ tensile deformation and hardness indentation studies have been carried out
separately on polycrystalline and single crystal materials. Baker et al. [100, 101] have
found that slip in NiAl is wavy and the ease of cross-slip leads to the formation of
dislocation cell structures at higher strains during in-situ tensile deformation of
polycrystalline NiAl. Morris et al. [102] have observed that only <001> dislocations are
mobile in a variety of single crystals. In binary NiAl and NiAl-Hf alloys they have
observed that a majority (70%) of the dislocations correspond to <100>{110} slip

systems. In NiAl-Fe alloys, only 30% of observed dislocations are <100> type and the

32

majority it

 

<1 i l> disli
found to be

Hot.
purity and
which dislrl
affects the -

ductility or

 

 

yet to be fu
The
in order to
would be .
degrees of

Of Single 1

comPositic

majority (65%) are <111> type dislocations. In contrast to <100> mobile dislocations,
<111> dislocations, which are activated near crack tips as a consequence of straining, are
found to be immobile.

However, none of these studies have directly observed the influence of alloy
purity and strain aging on dislocation generation and mobility in NiAl. It is not clear
which dislocations are susceptible to strain aging or in what manner the strain aging
affects the ability to generate mobile dislocations. The role of pre-stressing in achieving
ductility or generating mobile dislocations that are immune to subsequent strain aging is
yet to be fully understood.

The present study would use two different purity levels of the NiAl single crystals
in order to understand the role of impurities on dislocation behavior. The single crystals
would be cooled with different rates from elevated temperatures to induce different
degrees of strain aging. Also, different stress axes would be used to assess the influence
of single crystal orientation. Dislocations would be characterized as a function of

composition, orientation and cooling rate.

33

 

The
(TEM) an
understand:I
crystals. Til

commercial

influence 0‘

 

the use of r
ditiierent de
the Cooling
Hig
While cont-
Entitties.
dim-re“Ces
single CWS
Ca
Single Cry:
SlL'face. T
Ci’FSIaI.
Carbon TC:
We ETOr

Rig 80 p¥

3.0 EXPERIMENTAL PROCEDURE

The present study, which uses conventional transmission electron microscopy
(TEM) and in-situ TEM approaches, has been undertaken to achieve a better
understanding of the dislocation processes critical in enhancing plasticity in MA] single
crystals. The effect of interstitial impurities was studied by comparing the deformation of
commercially/conventionally pure (CP) and high purity (HP) single crystals. The
influence of single crystal orientation on deformation processes was investigated through
the use of differently oriented single crystals. The effect of cooling rate in imparting
different degrees of strain aging was characterized by adopting different cooling rates in
the cooling cycle.

High purity single crystals were provided by Dr. V. I. Levit, University of Florida,
while conventionally pure single crystals were provided by Dr. R. Daroila, G. E. Aircraft
Engines. Chemical analyses* of the as-received single crystals revealed significant
differences in the carbon concentrations between high and conventional purity NiAl
single crystals (Table II).

Carbon concentration varied to a large degree within the same sample in the CP
single crystals. A concentration as high as 272 ppm was recorded on a sample near the
surface. This may not be a true representative of the bulk carbon content of the single
crystal. Probably the EDM fluid used for cutting the chemical analysis sample left
carbon residue on the surface. Later on, all the surfaces in contact with the EDM fluid
were ground before performing the chemical analysis, and the resulting average value

was 80 ppm. No appreciable differences in silicon or oxygen levels were found between

34

 

m

l. Single t

 

g Commerc

 

Highl

\

 

t.AtIléil}."Sis

Table II. Impurity Concentration of the MA] Single Crystals

 

 

 

 

Single Crystals Si (Wt%) O (ppm) C (ppm)
Commercial Purity <0.001 380 79 - 272
High Purity <0.001 160 32

 

 

 

 

*Analysis was carried out at National Scientific Laboratory, (1-800-497-6752),
Cleveland, Ohio. For carbon analysis, a Leco carbon/sulfur deterrninator (Model CS-444
LS) was used, while a Leco nitrogen/oxygen simultaneous deterrninator (Model TC-436

DR) was used for oxygen.

35

 

the comm c

 

3.1. Then?

Single
a high-spe-
commercia
Centorr vac
alumina crt
requiremcn

during annq

 

mechanical
following tl
Within the f
Were 0.1 K
€001ng (FC

A t
reheated to
[Lindbetst
sample, M
cooling rat.
within the
With Hack

The mama

the commercially pure and the high purity single crystals.
3.1. Thermal Treatments

Single crystals were cut into rectangular pieces of 6.5 mm X 6.5 mm X 20 m using
a high-speed diamond-blade wafering saw. To homogenize the single crystals, both
commercially pure and high purity NiAl single crystals were annealed at 1600 K in a
Centorr vacuum fumace for 48 hours. The single crystals were placed in a recrystallized
alumina crucible for heat treatment. The rate of heating was adjusted with the vacuum
requirements of the furnace used for the homogenization anneal. The final pressure
during annealing was 10'6 mm of Hg. In order to study the influence of cooling rate on
mechanical behavior, different cooling rates were imposed on the single crystals.
following the homogenization anneal treatment. Accelerated cooling was accomplished
within the firmace using flowing argon. The imposed cooling rates on the single crystals
were 0.1 K/s and 1 K/s and are subsequently referred in the text as slow or furnace
cooling (F C) and fast or air-cooling (AC) respectively.

A few selected samples of commercially pure material were subsequently
reheated to elevated temperatures (473 K and 673 K) in a resistance type muffle furnace
(Lindberg). The samples were contained in an alumina crucible with a covered lid. The
samples were cooled inside the fumace (FC) and out of the furnace (AC) to vary the
cooling rate. For air-cooling, the samples were placed outside of the fumace but still
within the alumina crucible. The heat treatment schedules were adopted in accordance
with Hack et al. [18, 19] to have embrittled (FC) and toughened (AC) single crystals.

The thermal treatment process is schematically illustrated in figure 9.

36

Reheat

and

)

 

Flgum 9. :
Hack et al.

 

1600 K homogenized anneal for
48 hours in a vacuum atmosphere

7

 

 

 

Air-cooled Furnace cooled

   

 

 

 

 

Reheat to 673 K
And hold for 12 h

 

 

 

   

 

 

r Reheat to 473 K

 

 

 

 

 

 

 

 

 

 

 

473 PC 1600 AC 673 AC 673 PC 1600 PC

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 9. Schematic representation of the thermal treatment process in accordance with
Hack et a1. [18, 19]

37

3.2 Coml1r

Sell
were subjc
were grour
0.3 pm Al,
electrolytic
was central:
crystal bar
liquid nitrrl
continuousl

found to be

One

 

diffraction t
“40 it)" an.
<001>! and
3'3 C0mpre:

The
I“literature
hence“. C
CTillie initj;
inadi

ng Orb

to an apprg

apprOxlmate

3.2 Compression Sample Preparation

Selected specimens from both commercially pure and high purity single crystals
were subjected to compressive deformation after thermal treatments. Single crystal bars
were ground through 600 grit using SiC grinding paper and fiarther hand polished using
0.3 um A1203 polishing media. These bars were then electropolished at 243 K using an
electrolytic solution of one part of nitric acid and two parts of methanol. The solution
was contained in a stainless steel beaker, which was used as the cathode, and the single
crystal bar was used as the anode. The temperature of the bath was maintained by adding
liquid nitrogen periodically on the top of the electrolyte, which was being stirred
continuously by a magnetic stirrer. The applied potential for optimum polishing was
found to be 12 V.

Orientations of the single crystals were determined by back-reflection Laue X-ray
diffraction technique using a Seifert X-ray device operated with an acceleration voltage
of 40 kV and tube current of 30 mA. Single crystals were oriented to ‘soft’ <110>, ‘hard’
<001>, and some arbitraryorientations within the stereographic triangle (figure 10).

3.3 Compression Deformation

The polished and oriented specimens were then deformed in compression at room
temperature using an Instron (Model # 4206, Serial # 630) load-flame with a 100 kN
load-cell. One-inch diameter steel compression pads were used and precaution was taken
during initial contact of the compression pads with the sample to avoid asymmetric
loading or buckling. The applied crosshead speed was 0.06 mm/min, which corresponds
to an approximate strain rate of 4.4 X 10'5 /sec. The specimens were deformed to

approximately 2.5% plastic strain to induce dislocation generation and slip.

38

 

001

 

001 (Hard) 011 (Soft)

Figure 10. Orientation of the specimen within the stereographic triangle

39

3.4 TEM

At
technique
operated a
normal it:
cutting 3
parallel to
abrasive p

adjustable

 

Thr
jet polisher
melhanol u
A. The pu
3i“ COUrsr
foils as th
tilinning 0]
Storing for

Tra
“-800 (Se
200 W. I
dislocéition
in Order to

Primary Sp

3.4 TEM Sample Preparation

At this point the bars were cut into 3 mm diameter rods by spark-erosion
technique using a Materials Science NW MK 2 Electro Discharge Machine (EDM)
operated at 200 V. The rods were cut using a tube in an inclined way i.e. the slip plane
normal was made parallel to the cylinder axis. Thin foil specimens were prepared by
cutting 3 mm diameter discs from the rod by wire-EDM. The slices were sectioned
parallel to the slip plane. The discs were then thinned by hand grinding using SiC
abrasive papers to a thickness of approximately 200 pm. A steel grinding-ring with
adjustable depth was used as an aid for thin sample grinding.

The specimens were subsequently electropolished at 243 K using a Struers Twin
jet polisher (Tenupol 3) in a solution of freshly prepared one part nitric acid and two parts
methanol with an applied potential of 12 V and a resulting current of approximately 0.15
A. The pump speed was controlled during electropolishing by gradual decrease from 7 to
3 in course of the polishing. This precaution was taken to minimize any damage to the
foils as the foils became thinner. The specimens were cleaned thoroughly after the
thinning operation by fast dipping and rinsing in methanol at least three times before then
storing for microscopic observation. .

Transmission electron microscopic observations were carried out using a Hitachi
H-800 (Serial number 800-17-02) microscope operated with an accelerating voltage of
200 kV. Dislocations were observed using brightfield and weak-beam imaging. Initial
dislocation Burgers vectors were determined using the gob = 0 and gob x u = 0 criteria.
In order to perform a complete contrast analysis, all of the available g-vectors from the

primary spots of three poles, commonly {001}, {110} and {111} were utilized. Two

40

beam CO
reflection

dislocatio

 

3.5 lmagc‘

Di
dislocatio:
conditions

weakest ci

further cor

 

and true
determinin

using the c

beam conditions were set up with regular lattice spots excluding any superlattice
reflections and thus avoiding any further complications in interpretation of broad
dislocation images with large extinction distance.

3.5 Image Simulation

Due to high elastic anisotropy of NiAl, all of the planes around an edge or screw
dislocation are distorted; it was not always easy to achieve perfect invisibility even under
conditions of gob = 0 and gob x u = 0. Thus, the invisibility criterion often used as the
weakest contrast rather than complete invisibility of the dislocations. The situation gets
further complicated as most of the dislocations are neither pure edge nor pure screw type
and true invisibility is very difficult to achieve experimentally. Ambiguity in
determining the Burgers vectors of the dislocations was resolved by image simulation
using the computer program developed by Head et al. [103].

The computer program is based on the dynamical theory of image contrast and
essentially an image matching technique. A series of dislocation images are generated
and these simulated images are matched with the experimental image of any dislocation.
Following is the method of identifying the dislocations. Experimental micrographs of
the unknown dislocations are taken under different diffracting conditions i.e. different ‘g’
vectors taken in different electron beam directions. For accurate simulation, micrographs
and their corresponding diffracting patterns need to 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. It is important to recognize the
Kikuchi line patterns for the low index reflections. Important data includes diffracting

vector g for each micrograph, the beam direction B, the dislocation line direction It, the

41

foil norr.‘

Methods
|

lr.
images ;
observed
microgra;
with dit‘t‘e
such com
port of his
3.6 [rt-silt;
rd
machining

tensile 3pc

 

 

Wide slicel
00”esponr
Optimized
0f legs tha;
was foun d
Th.
kept little
103% and t
Then the .

J
4-

electrode it

foil normal F, foil thickness t, and a value of the deviation from the Bragg condition, w.
Methods of determining the above parameters are followed from Head et al. [103].

Informed guesses are then made about the unknown dislocations and theoretical
images are computed under the same diffracting conditions as the experimentally
observed image. The defect is identified when the theoretical and experimental
micrographs match consistently under different imaging conditions. Dislocation images
with different topological characteristics or distinguishable features are most suitable for
such comparison. This part of the work was carried out in conjunction with B-C. Ng as a
part of his Master’s thesis [104].
3.6 In-situ Tensile Straining

TEM tensile samples for in-situ straining were fabricated by electro discharge
machining to conform to a Hitachi H-5001 T tensile stage. A schematic diagram of these
tensile specimens is shown in Figure 11. At first a 1 mm thick, 15 mm long and 5 mm
wide slice was cut by wire-EDM, making sure that the tensile axis of the specimen
corresponds to the specific orientation of the single crystal. The initial thickness was
optimized for subsequent handling and machining. It was found that a starting thickness
of less than 1 mm was not able to sustain handling and some non-uniformity in thickness
was found from One end of the sample to the other after grinding.

The starting thickness for subsequent preparation and mechanical grinding was
kept little more than 1 mm. The gage lengths of the samples were approximately 3 mm
long and the gage-area was first created by Electro Discharge Machine profile cutting.
Then the 2 mm diameter loading holes were created by using a copper wire-rod as the

electrode in the EDM.

42

TCIBilt

 

Flgm'

Hole for Loading Pins

Tensile Axis t 6% é—_’ TEM single

tilt axis

 

 

 

 

 

Figure 1]. Schematic of the 15 mm long and 5 mm wide in-situ tensile specimen

43

rounded

specimen
I

using Sit

mentionci
El
tensile Spi

methanol

 

 

mentionec
generating
Were Scray
Tenupol 1
dining lor
of the 5p.
methanol

ObSerVati

The sharp edges at one end of the specimen (the left-hand side of figure 14) were
rounded by hand grinding to fit in the tensile stage of the microscope. The tensile
specimens were subsequently hand ground to a final thickness of approximately 150 um
using SiC abrasive paper. Masking tape and double-sided tapes were used to assist
grinding. The tape was wrapped on the fingers before grinding to prevent slippage of the
samples. The specimens were subjected to a quick electropolish to minimize the effects
of any surface flaw or contamination due to grinding using the same conditions
mentioned in section 3.2.

Electron transparent thin areas were then created within the gage length of the
tensile specimens by careful twin-jet electropolishing in a freshly prepared solution of
methanol and nitric acid (2:1) following the same steps, equipment and conditions
mentioned in section 3.4. Sometimes the specimens slipped during electropolishing
generating a hole that was not located centrally within the gage length. These specimens
were scrapped. As the gripping of the tensile specimens within the sample holder of the
Tenupol twin-jet polisher is essentially by mechanical contact, precautions were taken
during locking the two male-female parts of the sample holder to keep only the gage area
of the specimens exposed to the electrolyte. The specimens were thoroughly rinsed in
methanol (as mentioned in section 3.4) following electropolishing prior to TEM
observation.

In-situ straining experiments were carried out in the same Hitachi H-800 TEM.
The drive for azimuthal rotation was utilized for tensile straining. The whole area of
observation was mapped prior to any straining. The straining was started very slowly

until any noticeable dislocation motion was observed. Dislocation activity and the

mobile c.
wide-an;

HS-U28‘t

 

straining =
due to in
areas of ii

Sr
length an
failure of

Lt

 

diffracting
dining in-I
not possil:
Qm'ed or
motion,

dlSlocatio
bOYIOm O
dll’E‘CtiOng

Plane of‘

(1.0

'71'l-sit
Ir
diSIOCarri

COuld 0n

mobile dislocations generated by in-situ straining were recorded dynamically using a
wide-angle CCD camera (Gatan 673-0200) and a VHS videocassette recorder (Mitsubishi
HS-U28), while static images were recorded with the TEM plate camera. Further tensile
straining was continued until the mobile dislocations were observed to cease their motion
due to interaction with pile up of dislocations. Mobility of dislocations in the thicker
areas of the specimen was also difficult to follow and record.

Sometimes cracks opened up from the thinning hole in the middle part of the gage
length and the specimens failed during straining. Attempts were made to avoid the
failure of the strained samples for subsequent static observation.

Unfortunately the TEM tensile stage has only one tilt axis available to vary
diffracting conditions, making it impossible to obtain variety of diffracting conditions
during in-situ straining. Thus, complete characterization of the mobile dislocations was
not possible. However, limited analysis of dislocation line directions and slip planes was
carried out based on slip trace analysis of the observed slip lines following dislocation
motion. Dislocation motion was evidenced with slip trails left in the wake of the
dislocations. These trails are the consequence of the dislocations disturbing the top and
bottom of the surface oxide layer on the foils as they pass. Inference on the line
directions of dislocations was based primarily on the dislocations lying normal to the
plane of the foil.

3.7 In-situ Tensile Deformation of Disc Samples

In order to determine the Burgers vectors and line directions of the mobile

dislocations, it was necessary to utilize a wide variety of diffracting conditions, which

could only be accomplished by using a double-tilt specimen holder. Those specimen

45

holders '

   

requirenr
the gage
TEM. A
to mecha
the toy '
parameter
use a diar
kept at so
the sampl
\V
Specimen

SP‘v‘k‘imens

 

Control thd
3 aVOidir
alinost irr
mlfrest of
damage to

Fit
help of 10
ill

Uh 306cm

llllS way l

holders are designed to accommodate 3 mm diameter disc specimens only. To meet the
requirements of the double-tilt specimen holders, several attempts were made to extract
the gage-length section of the 15 mm long specimens after tensile deformation in the
TEM. At first, tube EDM was used to cut a 3 mm diameter disc from the gage area. Due
to mechanical vibration involved in cyclic pulse, the thin areas fractured and were lost in
the EDM bath during spark erosion. All kinds of manipulation with the operating
parameters could not retain the electron-transparent part within the disc. Next attempt to
use a diamond core-drill was also unsuccessful, as the mechanical vibrations could not be
kept at such a low level so that the thin section remained adhered with the main-body of
the sample.

When all those attempts to remove the deformed thin areas from the tensile
specimen failed, an alternate approach was taken to deform a 3—mm disc specimen. Disc
specimens were prepared and glued to the tensile stage. Difficulty was experienced to
control the amount of glue to ensure enough mechanical strength during straining, as well
as avoiding masking the electron-transparent section by the glue. In addition, it was
almost impossible to unglue the deformed specimens without damaging the areas of
interest of the test-piece. Use of solvents, like acetone, to dissolve the glue also caused
damage to the foil.

Finally, 3-mm disc specimens were prepared and mechanically clamped with the
help of loading screws in the tensile holder and strained in tension within the electron
microscope. The same samples were subsequently observed in the TEM using the double
tilt specimen holder keeping in conformance with the positioning in the tensile stage. In

this way the mobile dislocations were easily characterized in relation to the dislocations

46

generals
ai'ailabh

criteria t

 

 

generated during tensile deformation. Thus, a fill range of reflecting conditions were
available and complete contrast analysis was performed using g-b = 0 and g-b x u = 0

criteria to determine the line directions and Burger’s vectors of the mobile dislocations.

47

The met
obtainet
fracture
crystals.
are pres

Ni-Al si

Single c
Crystal r
induced
t6imperat
followed
and dish

T
Statican}
CCTV C
SOme re
West-ma
examplfi

complete

 

4.0 EXPERIMENTAL RESULTS

The results of the experiments outlined previously are presented in this section.
The mechanical behavior of the single crystals is presented first. Yield stress properties,
obtained by compression tests of the commercially pure materials, are compared with the
fracture toughness properties determined by Hack et al. [18, 19] for similar single
crystals. Then the results of post-deformation or static transmission electron microscopy
are presented before going into the details of in-situ transmission electron microscopy of
Ni-Al single crystals.

The evolution of dislocation substructures in commercially pure undeformed
single crystals is presented as a function of thermal treatments. Undeforrned single
crystal results are followed by results from deformed specimens. The deformation
induced dislocation sub-structures are presented in relationship to the reheating
temperature and variation in cooling rate. Commercial purity single crystals results are
followed by the results from the high purity materials where the mechanical properties
and dislocation sub-structures derived from compressive deformation are presented.

This post mortem study is followed by in-situ tensile deformation, where
statically captured images for dynamic movement of dislocations are presented. A
CCTV camera was used to record the dislocation motion at the onset of deformation.
Some representative clippings (with a span of 64 minutes) from these recordings are
presented in the form of digital video disk (DVD) and included in the Appendix. A few
examples of dislocation image simulations are included also for the sake of achieving

completeness in dislocation characterization. The last part of the results includes the in-

48

rim dc.

 

m’SlfllS

using a

values a'
values oi

[18.19].

 

indicate t
in liactur
in a con
intermed
dlll'lblllfi
Discussir

hiatus: t

situ deformation of 3-mm disc specimens in a single-tilt holder from high purity single
crystals and observation of the same post deformation dislocation substructures in TEM
using a double tilt specimen holder.
4.] Mechanical Behavior
4.1.1 Compression deformation of commercially pure single crystals

The results of the compression tests are presented in Table III. The yield stress
values at different thermal treatment conditions are compared with the fracture toughness
values of similarly treated commercially pure NiAl single crystals obtained by Hack et al.
[18, 19]. Figure 12 shows the trends of these two mechanical properties. The results
indicate that slower cooling rates increase the yield stress with a corresponding decrease
in fracture toughness. Conversely, faster cooling rates decrease the yield stress and result
in a concurrent increase in fracture toughness. Reheating the single crystals to an
intermediate temperature revealed significant impact on the final properties. This may be
attributed to the so-called strain aging effect that will be discussed at greater length in the
Discussion chapter (5.0). It is also observed that, irrespective of the previous thermal
history, the final cooling rate strongly influences the mechanical properties.
4.1.2 Compression deformation of high purity single crystals

Table IV lists the yield-stress values for fumace-cooled and air-cooled high purity
NiAl single crystal samples. These single crystals were not treated at intermediate
temperatures like commercial purity materials as the influence of thermal conditiOns
especially cooling rate was not that significant in altering the tensile properties. The
yield stress values, determined by compression tests for high purity single crystals, are

lower compared to similarly treated commercially pure material.

49

Table l

 

l 600

673

 

 

Table III. Comparison of yield stress obtained by compression test of commercially pure
NiAl single crystals with fracture toughness. *

 

Yield Fracture
Specimen Thermal History Stress Toughness

(MPa) MPa\/m*

 

473 FC 1600 K homogenization anneal 400+/-12 2.8
and fast-cooled.
Reheated to 473 K and slow-cooled.

 

1600 AC 1600 K homogenization anneal 200+/-8 15.6
and fast-cooled.

 

 

673 AC 1600 K homogenization anneal 153+/-5 16.7
and slow-cooled.
Reheated to 673 K and fast-cooled.
673 FC 1600 K homogenization anneal 361+/-6 5.8

and slow-cooled.
Reheated to 673 K and slow-cooled.

 

1600 PC 1600 K homogenization anneal 450+/-10 2.4
and slow-cooled.

 

 

 

 

 

 

*Adapted from Hack et al. [18, 19]

50

 

 

 

 

 

 

 

 

Yield behavior of commercially pure single crystals
500 18
450 . y . 16
IYteld stress
400 C] Fracture 7 1 4 A
Q
350 ~ 2—
A 12 g
N
D_ Q
2 300 E
8 10 §
g 250 E
4’; 8 at
1, 200 g
>- 5 g
150 T 6
9
LL
100 ~ 4
so 2
0 0
473 F0 1600 AC 673 AC 673 F0 1600 FC
Condition

 

 

 

Figure 12. Change of yield stress and fracture toughness (Per Hack et al. [18, 19]) with
thermal history. The inverse relationship of yield stress and fracture toughness may be
noted.

51

Ta

 

 

Table IV. Yield stress properties obtained by compression test of high purity NiAl single

 

 

 

 

 

crystals
Yield Stress
Specimen Thermal History (MPa)
1600 AC 1600 K homogenization anneal and fast-cooled. 142+/-10
1600 K homogenization anneal and slow-cooled.
673 AC Reheated to 673K and air-cooled. 135+/-7
1600 K homogenization anneal and slow-cooled.
673 PC Reheated to 673K and slow-cooled. 274+/—9
1600 PC 1600 K homogenization anneal and slow-cooled. 287+/-5

 

 

 

 

 

52

4.2. Slip Trace Analysis

Slip lines were observed using optical microscopy and the active slip planes were
determined by slip trace analysis following the method described by Barret [105]. Figure
13 shows an example of slip lines developed on two perpendicular faces of a deformed
single crystal. The angles or and B were measured on the micrographs taken from the
front and right faces of the rectangular compression specimen as shown in Figure 14.
Then the angles, or and B, were plotted on the stereographic projection of the front face in
the way as shown in figure 15. The two points define the slip plane, which causes the
slip trace on the single crystal. The slip plane and its pole (90° from the plane) are shown
on the stereographic projection. Slip was observed on {001} planes for all of the
commercially pure specimens. The majority of the high purity specimens also exhibited
traces from {001} planes and in some cases {110} planes. There were a few cases when
limited cross-slip was observed and {112} was determined as the cross-slip plane.
4.3 Static/Post-deformation TEM Observations
4.3.1 Undeformed commercially pure single crystals

Undeformed specimens exhibited isolated and widely spaced dislocations
throughout the structure, with few dislocation loops and tangles. Many observed areas
had essentially no dislocations. For undeformed materials, both slow cooled and fast
cooled specimens had similar dislocation densities as is evident from figure 16 for 473

FC, 673 PC and 673 AC samples respectively.

53

 

Figure 13. Two faces of a deformed specimen showing slip traces with measured angles
or and B on (a) front and (b) right faces respectively.

54

Compression Axis

 

 

 

 

V
Top Face
Front Face R
(1 2'
Right Face
\

Figure 14. Schematic representation of compression sample showing the measurement of
or and [3 used in determining the slip plane.

55

 

 

 

Figure 15. Stereographic projection of the single crystal compression sample shown in
figure 13 displaying the plotting of angles or and [3 used to determine the slip plane.

56

 

Figure 16. Isolated, widely spaced, single dislocations are observed throughout the
structure of commercially pure NiAl single crystals with few dislocation loops and
tangles in (a) an undeformed 473 FC specimen. (b) an undeformed 673 PC specimen and
(c) an undeformed 673 AC specimen respectively.

57

 

0T

lllr

ml

C0

lllli

Dislocations were characterized using the invisibility criteria; gob = 0 for screw
dislocations and gobxu = 0 for edge dislocations. Depending on the values of g, these
factors can assume a zero value signifying no displacement field around a dislocation
line. A dislocation becomes invisible if there is no displacement field surrounding the
dislocation. These criteria hold good with complete invisibility for isotropic materials
only. The factor, gobxu for edge dislocations, can be zero only where the line direction u
is parallel to the diffraction vector g. However, NiAl is a highly anisotropic material and
dislocations are often mixed in character. Thus, dislocation lines may not go completely
out of contrast in many situations, even after satisfying the invisibility criterion.

The evolution of the dislocation structure was studied and the Burgers vectors of
the dislocations were analyzed using this invisibility criterion. In absence of complete
invisibility, the weakest contrast was used as an approximation for the out-of-contrast
conditions. A series of reflections were used to set up two beam conditions with regular
lattice spots. Two such reflections for invisibility were used to determine the Burgers
vector b, which is essentially the zone axis of the two reflecting planes.

Figure 17 illustrates an example of such dislocation analyses for slow cooled
undeformed 673 PC specimens. Two types of dislocations were identified as types ‘a’
and ‘b’. Type ‘a’ dislocations, which remained in good contrast for the [011] reflection
and visible with weak contrast for the [200] and [110] reflections respectively, were
determined to have the [001] slip direction. Similarly, ‘b’ type dislocations observed to

be in weak contrast for the [200] and [111] reflections, were determined to have the

[OT 1] slip direction.

58

 

Figure 17. The same area of undeformed commercially pure 673 PC specimen is shown
with different tilts giving g = [011], [200] and [110]; ‘a’ denotes [001] and ‘b’ denotes

[0T 1] slip directions.

59

All thermal treatments of undeformed specimens resulted primarily in <001>
dislocations with a few <110> dislocations. Some of the dislocations in undeformed
specimens were observed to move as a result of beam heating. Unlike deformed
materials, no tangle of dislocation was observed and mobility of such dislocations was
not constrained. Irrespective of the thermal treatments, {001} was analyzed as the slip
plane in all cases.

4.3.2 Deformed commercially pure single crystals

The gross morphology of the dislocation structures differed significantly between
deformed and undeformed specimens. Compression deformation resulted dislocation
tangles, which were not observed in undeformed specimens. Dislocations were also seen
to move as a result of beam heating which results in strains in the foil due to different
thermal expansion in different areas of the foil. In slow-cooled deformed materials
dislocation movement under beam heating conditions ceased as soon as the mobile
dislocations interacted with the pre-existing dislocation tangles, which were already
pinned by solute atoms during slow cooling through the elevated temperature region.
4.3.2.1 Slow cooled (FC) samples

Slow cooled samples, which exhibited higher yield stress than fast cooled
samples, were found to exhibit dislocation structures that were heavily tangled in
comparison to fast cooled samples from either the homogenization anneal temperature
(1600 K) or any intermediate reheating temperature (473 K/673 K). Figures 18 - 20
presents typical dislocation structures obtained through slow cooling of 473 FC, 673 FC

and 1600 FC samples that exhibited higher yield stress (2 400 MPa) values.

60

 

Figure 18. Heavily tangled dislocation morphology in deformed commercially pure 473
FC specimen; (a) brightfield image and (b) weak beam image.

61

 

Figure 19. Deformed commercially pure 673 FC specimens reveal tangled dislocations;
(a) brightfield and (b) weak beam image

62

 

Figure 20. Deformed commercially pure 1600 PC specimens display heavily tangled
dislocations in (a) brightfield and (b) weak beam image.

63

 

 

'1‘)

r"?

Figure 21 illustrates an example of dislocation analyses for slow cooled 673 K
samples. Two different types of dislocations were identified as type ‘a’ and ‘b’. Type
‘a’ dislocations, which remained in good contrast in (3), went in extinction for (b) and

visible with residual contrast in (c), were determined to have [100] slip direction.
Similarly, ‘b’ type dislocations, which were observed to be in weak contrast for [2 00]

and [2 11] reflections, were determined to have [01T] slip direction. All other deformed
slow cooled samples from 473 K and 1600 K were found to have both <001> and <110>
dislocations, the majority of which were <00] > dislocations. Dislocations were observed
to move under beam heating conditions in some cases.
4.3.2.2 Fast cooled (AC) samples

A significantly different dislocation substructure was observed for deformed fast-
cooled samples (AC), Cooled either from 1600 K or from 673 K. Analyses of the
dislocation structures are shown in figures 22 and 23 respectively. Specimens that have
undergone slow cooling tend to develop dislocation tangles, while fast cooled specimens
exhibited dislocations with wider separations and were sparsely entangled. In contrast to
slow cooled samples, dislocations were seen to move more fi'eely for fast cooled
specimens under beam heating conditions; this may be attributed to the less intense
interaction of the mobile dislocations and dislocation tangles.

Figure 24 illustrates an example of dislocation analyses for fast-cooled 673 AC
specimens that possess lower yield strength than the slow cooled specimens. In this case

two different types of dislocations were observed and identified as types ‘a’ and ‘b’.

Type ‘a’ dislocations displayed weak contrast in [010] and in [01 1] reflections and were

64

 

Figure 21. Commercially pure 673 FC specimen reveals presence of two different types
of dislocations, a with b = [100] and b with b = [011]. The same area is shown with

different operative reflections; g= (a) [T 10], (b) [01 T] and (c) [020]. Dislocations were
also seen to move in (c).

65

 

Figure 22. Dislocation structures observed in deformed commercially pure 1600 AC
specimens; (a) brightfield image and (b) weak beam image.

66

 

Figure 23. Very few dislocation tangles were observed in deformed commercially pure
673 AC specimens; (a) bright field image and (b) weak beam image.

67

 

Figure 24. Commercially pure deformed 673 AC specimen shows ‘a’ and ‘b’
dislocations with b = [i001 and [01 i] respectively. The same region with different
operative reflections are shown in (a) [l l0], (b) [200] and (c) [2i l].

68

determined to have the [l 00] slip direction. Type ‘D’ dislocations were observed to be in
sharp contrast for g = [l l0] (figure a), in extinction for g = [200] (figure b), and in

residual contrast for g = [2i l] (figure c); they were determined to have the [01 l] slip
direction. Similarly, deformed fast-cooled samples from 1600 K were also analyzed to
have both <001> and <1 10> dislocations.

Although the types of dislocations did not vary between slow and fast cooled
samples, the differences in the gross morphology of the dislocations of deformed
specimens are striking. Samples slowly cooled through the temperature window of 673
K to room temperature display heavily tangled dislocations. This observation suggests
the possibility of static strain aging, where interstitial atoms such as carbon had ample
time to diffuse to the core of the dislocations during slow cooling of the specimens.
Table V summarizes the types of dislocations observed in relation to the thermal history
of the specimens and their yield stress values.

4.3.3 Compression deformation of high purity single crystals

Post compression deformation dislocation substructures developed in the high
purity single crystals differed significantly from the commercially pure single crystals.
Only limited tangles of dislocations were observed. Dislocations were seen to move
under the influence of beam heating irrespective of the thermal treatments prior to
deformation. No discernible difference was observed between the dislocation behaviors
of the specimens depending on the final cooling rate. Dislocations generated during
deformation of high purity single crystals were analyzed in the same way as for

commercially pure single crystals.

69

Table V. Burgers vectors analyses for deformed specimens in relation to their thermal
histories and yield stress in commercial purity materials

 

 

 

 

 

 

 

 

 

 

Specimens Yield Stress (MPa) Burgers vector b
473 FC 400 Primarily <001> Scattered <01 l>
1600 AC 200 Primarily <001> Scattered <01 1>
673 AC 153 Primarily <001> Scattered <011>
673 PC 361 Primarily <001> Scattered <011>
1600 FC 450 Primarily <00] > Scattered <01 l>

 

7O

 

Figures 25 and 26 present dislocation structures obtained through fast and slow
cooling of 1600 AC and 1600 F C samples respectively. In both cases dislocations were
observed to move under the influence of the beam and were characterized in reference to
immobile dislocations. Figures 25 (a—d) reveals the mobile dislocations to be [001] type
for fast cooled samples. A similar situation was observed for slow cooled samples, where
mobile dislocations were determined to be of the [001] type, as shown in figures 26 (a-d).
Dislocations lying in the plane of the foil were also found to be [001] type.

Dislocation analyses from samples subjected to intermediate reheating

temperature of 673 K are presented as figures 27 and 28 respectively. Figure 27 (a-c)

show [I 01] dislocations for 673 AC or fast cooled samples. Although presence of <1 1 1>
mobile dislocations is unlikely in NiAl, figure 28 (a-d) reveals [111] as the mobile
dislocations in the 673 F C or slow-cooled samples. Overall the gross morphology of the
dislocations differed significantly between the deformed microstructures of high and
commercial purity single crystals, but there is not much influence of the cooling rate on
the dislocation morphology in high purity single crystals. In undeformed high purity
single crystals there were a few widely spaced isolated dislocations present; consequently
no images are presented due to lack of value added information.
4.4 Image Simulation

As mentioned in the literature review section, due to high elastic anisotropy of
NiAl it is almost impossible to obtain perfect invisibility of dislocations in MA] single
crystals. Therefore, the contrast analysis was complemented with image simulation.
This section of the results describes a typical image simulation for a dislocation in a

undeformed 673 AC specimen.

71

 

Figure 25. Dislocation analyses in high purity 1600 AC specimen shows [001]
dislocations in different reflections. Dislocations that are mobile under beam heating
condition appear with good contrast in figure (a) and out of contrast in figures (b), (c) and

(d).

72

ti ‘um
«....

 

Figure 26. High purity 1600 FC specimen shows [001] mobile dislocations (marked as
a) in different reflections. Type b dislocations, which are not moving due to the effect of
beam heating, also seem to have similar Burgers vectors.

73

 

Figure 27. Dislocation characterization in high purity 673 AC specimens shows [101]
dislocations in difi‘erent reflections. Dislocations, which are almost parallel to the beam
direction in figure (a), are characterized with reference to several immobile dislocations
present in the microstructure. Such mobile dislocations are out of contrast in figures (b)
and (c).

74

 

Figure 28. It is not usual to find [111] mobile dislocations in NiAl. Figures (a) and (c)
show such dislocations in good contrast whereas mobile dislocations are observed with
very weak and residual contrast in figures (b) and (d) in high purity 673 F C specimens.

75

An image of dislocations in the 673 AC undeformed commercially pure specimen
is shown as figure 29. The dislocation marked ‘A’ is representative of most of the
dislocations present in the micrograph. For better comparison with the computed image,
several experimental images of the same dislocation were captured using three non-
coplanar diffracting vectors.

Dislocation images from different diffracting vectors are helpful in obtaining
information about all components of the displacement field of the defect. Figure 30
shows a set of seven experimental images of the dislocation ‘A’ with different diffraction
vectors (a-g) tabulated in Table VI. Four sets of computer generated images for the same
diffraction conditions and known Burgers vectors are also shown in the same figure for
comparison. Important parameters to obtain those simulated images are also listed in
Table IV; these parameters are diffracting vector g, beam direction B, dislocation line
direction u, foil normal F, foil thickness t, and deviation from the Bragg condition, a).
The object was to generate a set of computed images and find the best visual agreement
with the experimental set.

It is evident from the computed images that the [100] and [110] Burgers vectors
are not in agreement with the experimental image, especially when the dislocation in (c)
goes close to invisibility or at its weakest contrast. The Burgers vector of [001] might be
a possibility, but the computed image fails to match with the experimental image (d) for
contrast. Based on the fact, these three Burgers vectors for dislocation ‘A’ may be ruled
out. Further comparison of the remaining computed and experimental images confirms
this conclusion. Moreover, visual agreement between experimental and simulated images

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

76

 

Figure 29. Dislocation ‘A’ in an undeformed commercially pure 673 AC specimen used
for image simulation and generation of computed image.

77

.552: $985 EowSm on. 8 wemveoamoboo 802280 eouogme 08mm 05 c8 woman: 339:8 we 33 58 can
_> 038. 5 53m macaw—Eco nouofibE 05 55 :83 .<. aoaaooiv mo 3 - A8 8925 808593 .«o “om < .3 8:me

SE

    

., . r are

e .n.

.aglrfilwrzrn , ..lrormttt. a

‘ ~....
w

 

 

 

. . . ..
#3.... .. .l ........)..r..L. .oen.e..§.&.t.rr.. 2.659.

on. . r. ....Jnflogqmuizflfw... ....r...
....s . fervor. . . .

Hf. . ... ....
... is. .

1......

 

 

Mun. 1....r...... .........l.... i .. ... a . Hind
_wu

 

_ .7?"

'--
I
l‘

 

 

-5:

uJ...2..E.1....M ..u......»u.....t......1....:.w..r. satin: vita...“
.. . . . .

[£33. H.

1!"
if .
i W‘W‘I‘ in.

 

 

 

 

 

    

 

9......m..11..:....... ....3....r_n.......i.. ..a HR: .

 

 

a _

.
.... .. .. ......n... .. .. .
F...t.r..:fc..z. ...eite...¢..t$:fi~.‘...e .L..:...”..... ....k

....w.......4......m.....“no.........u...\.......34....uihi.q.......a._iad
... ......n
e. _. . .
m .

Ehwevrnaezutfifiefib. .EVuasruufirw

 

 

. . ... ..... ... . ....«rn .. .... . Fiat... . .... ...”..‘fi
.. . ., . . ..
.i - svv. on. 4 . . . r .y
. .J
....
. .
. .

 

:oa

. tmflifi Nauxnmnralfa... p

 

”an... ......JQ... fie1~uflmxuz.3....d:;u...s teeth;
. . . . .

W __...uv Q Is In“

M
m .
m.

Ifl'"“ -

I

grantefix r... “EV... rice _gfir‘wL

i ..

w

.v

3
.u.
W.

 

 

W Swanson... “Miamia..u.,..,.__...1w...r..3.33.

a . . M.

v k. . . .
Nu...

,Q t. (€328)ng \e’t .
I‘4

.

.rV . . . . . ..
N ‘ ”ii-v». ... ,I..LF .. . Ii... . innit... ....»i e . uh.“

loo-Iz‘lluvaeti flu-iii; 4|» .... n. e s
. . .. . . _ . . . .n . .
. . T . . .. . . . .

 

“ox
r.
w...
M
u...
.5

trike.-.nit...merrier..."......r..:.i.......\.... . .

........ ...1... . ,....
.3..1....:... ......rse..e.-c..1e..u ~2......fiv.2.....; ....L

 

 

owns: Equatogxm

«a ... ”queue- anaemia... harmony raj

.-

T _ . .

.
p
. F
F .. tigtrrgrisullt “sun?”

3

.. 3...... .3 3,345.13... $1.33!...

  

“far:- 1'???“- ‘~
3.
Q
I
1
h
3
i
E
O
P
i
s
Q.
_ ~ "

. . ..Q
3:13;]: reciting-fl

r “mafia—rah rarefiie Q: Jessi... .
. . w .. . . ., x . . . . .e A. ‘
. . . a.
. v . ...
_ .. .h .. __..
.... v . . ......
, ....r tart; . ... ...

. . . . .... v. . ~ . V
U I . v . .I.

. e . ... . . v‘...

..n. .....m....!- .ICF. Brief 2.4;» r. . um...r.h.r.Li.-.m .

. ...... i. .. ..
.. .. fl... _ A .1...

t

 

’1‘!
.i
'1

I

Q

r

3‘.

a .. «.-.-w--—

   

~\.. -»

u. .... . .. A. . . .... .
Mn. {I ,. . L}. ... L e \L .
I.

‘nL' .". ' 3‘ h '

an...” 2 . ...........3.

2

 

78

Table VI. Parameters used to simulate dislocation ‘A’

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a b c d e_ - -f g
g ” 5” 0 0. " " " 110 ’ ‘ 0 2.6 200 "”011 T10" “01'3“
11 3236 3236 3236 3236 3236 3236 3236
B 016 2215 2019 0911 T33 i515 299
F i112 i112 1112 i112 i112 ’1112 i112
or 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,

g = diffracting vector

u = dislocation line direction

B = direction of electron beam

F= foil normal

60 = deviation from Bragg condition and this is the dimensionless deviation parameter

t = foil thickness in extinction distances

Other parameters used for the simulation include the following:

Elastic constants: C 11 = 2.03

and

C 12: 1.34

Anomalous absorption coefficient: ANO = 0.07

79

C44= 1.16

 

4.5 In-situ Transmission Electron Microscopy

To study the dislocation motion in both high and commercially pure NiAl single
crystals, 15 mm long micro-tensile specimens were deformed in the tensile stage of the
TEM. In all of the samples using in-sr'tu straining, dislocation motion was readily
observed, regardless of the impurity concentration, orientation, and thermal history.
Figure 31 presents a series of photographs depicting a sequence of dislocation motion
resulting from in-situ straining. Prior to the in-situ straining, the morphology of the pre-
existing dislocations was examined in each case and the samples were mapped for
possible reference points to check any movement of the dislocations. Figure 31 (a)
represents an area with pre-existing dislocations (PED).

Motion of dislocations was typically characterized by clearly evident slip traces
left in the wake of dislocations. In most cases, dislocations were generated at the foil
edge and slipped into the material. Dislocations were also generated from sources deep
within the foil (it was not possible to observe these sources, as the foil was too thick to
observe) and moved in the tensile direction (figure 31 b). These dislocations, lying
perpendicular to the foil, moved in slip bands and slipped through the pre-existing
dislocations lying in the plane of the foil. With straining, dislocations were seen to move
by rapid jumps towards the edge of the foil (figure 31 0). While the dislocations tend to
move in relatively well-defined slip planes, some limited cross-slip was observed as
evidenced by waviness in the slip traces. Because the slip traces are presumably surface
steps, the end-on dislocations must have a screw component. Motion of dislocations
lying nearly perpendicular to a foil will only leave surface steps if the dislocations have a

screw component.

80

 

Figure 31. Dislocation motion in high purity slow-cooled [110] oriented NiAl. Sequence
of events captured as (a), (b) and (c) revealed progressive movement of dislocations in
the direction of the arrow. As a result of straining dislocations generated within the foil
and starts moving towards the edge.

81

 

Figure 31 (contd.). Dislocations, almost perpendicular to the foil moved in the slip bands
and cross-past pre-existing dislocations in the plane of the foil. Sequence of events (d),
(e) and (f) revealed slight waviness in the slip bands that indicates cross-slip of such end-
on dislocations. Figures (e) and (f) show development of dense slip bands.

 

Figure 31 (contd.). Eventual crack initiation from the edge of the foil along the slip
bands that produces avalanche of dislocations at its tip but the pre-existing dislocations
stays sessile throughout the process.

83

Dense slip bands developed from the foil edges with further straining (figure 31
(1). Within these slip bands, significant debris accumulated, resulting in dislocation
pinning and loop formation. In the later stages of deformation, cracks developed at the
foil edge and propagated along the slip bands, (figure 31 e through i), while dislocations
piled up in the slip band ahead of the crack-tip.

Unfortunately, it has not been possible to relate the observed motion to
macroscopic stresses and strains. This is because foil-to-foil variation in thickness and
stress-concentration cannot be quantified in an accurate manner. The mobility of the
dislocations and the dislocation activity at the onset of tensile straining were recorded
dynamically and presented in the annexure as a digital video disc (DVD). Some static
images were captured in between straining steps and presented here.

Dislocation velocities were also calculated. The values of the dislocation
velocities provided a basis for qualitative comparison of dislocation mobility between the
high and commercial purity single crystals. It should be noted that the absolute values of
the dislocation velocities might be severely influenced by thin foil effects. Also, it is
essentially impossible to relate the imposed strain rate of the tensile stage to the
dislocation velocities observed in the complex state of stress generated around the
thinning hole in the TEM foils. However, the relative difference in the velocities
between the materials indicates significant differences in dislocation mobility. Thus, the
measured velocities (with the constraints of the experimental conditions) reported in
Table VII are used solely for the sake of comparison of dislocation behavior as a function

of impurity concentration and cooling rates.

84

Table VII. Dislocation velocities measured at the onset of straining using CCTV camera

 

 

 

 

 

 

 

 

 

 

Tensile Dislocation
Specimen Thermal History axis/ Velocity
Orientation (urn/s)
Commercial 1600 K homogenization anneal and
Purity fast-cooled <100> 0.1 - 1.4
Commercial 1600 K homogenization anneal and
Purity slow -cooled <100> 0.5 - 2
Commercial 1600 K homogenization anneal and
Purity fast-cooled <110> 1.5 - 3
Commercial 1600 K homogenization anneal and
Purity slow-cooled <110> 1 - 3
1600 K homogenization anneal and
High Purity fast-cooled <100> 5 - 25
1600K homogenization anneal and
High Purity slow-cooled <100> 4 - 20
1600 K homogenization anneal and
High Purity fast-cooled <110> 10 - 25
1600 K homogenization anneal and
High Purity slow-cooled <110> 13 - 17

 

 

 

 

85

 

Despite the significant differences in the dislocation velocities, it should be noted
that as the deformation substructures developed, multiplication of the new dislocations
within the slip bands was observed in all of the materials. Results from both commercial
purity and high purity single crystals will be presented in the following sections. Some
observations from the 3 mm disc straining of high purity specimens will be included in
the last part of the results.

4.5.1 Commercial purity specimens

Dislocation motion studied in commercial purity specimens will be presented
first. Tensile specimens were oriented such that the tensile axes of the specimens are
near the ‘hard’ <001> or ‘soft’ <110> orientations of the single crystals. Studies were
also conducted with specimens oriented to some arbitrary tensile axis such as [23 8].
Prior to in-situ straining, the morphology of the pre-existing dislocations was examined.
Figure 32 shows a typical commercially pure sample with tangles of dislocations in slow-
cooled condition. Many of the dislocations are extensively jogged and a few dipoles are
also observed. Planar arrays of dislocations, which were observed in many parts of the
foil prior to straining, remained immobile during the deformation.

Commercially Pure Air-Cooled [23 8]

Figure 33 shows dislocation activity in course of in-situ straining of commercially
pure specimens. No movement of dislocations was noticed at the very start of the
straining process. There may be some slop in the loading hole, which delayed the actual
load on the specimen. The dislocations originated primarily from the edge of the foil
with straining, and slipped in the tensile direction as well as at an angle to it. Distinct slip

bands are observed in figure 33 (a). Figures 33 (b) and (c) show progressive movement

86

 

Figure 32. A representative microstructure of commercially pure NiAl single crystal
exhibit tangles of dislocations. Many of the dislocations are extensively jogged and a
few dipoles are also observed.

87

1;: m

 

Figure 33. (a) Bright field image illustrating the nucleation of mobile dislocations at the
edge (bottom left hand comer in the micrograph) of the foil in commercially pure air-
cooled NiAl oriented to [23 8]. (b) and (c) Images showing the dislocation motion in two
S] p bands displaying straight dislocation trails indicating no cross-slip. (d) Eventual
cracking from the foil edge, propagating between slip bands.

88

of mobile dislocations slipping in two bands that developed from the same part on the foil
edge. In the commercially pure samples, slip was very planar and the dislocations moved
at a fairly constant, slow rate, rather than through quick jumps observed in high purity
materials. Once significant slip occurred, cracks developed between the slip bands and
punched out more dislocations at their tips (figure 33 (I). There is some build up of loop
debris in the slip bands afier extensive straining. The line direction of the dislocations
was near [100] in both of the slip planes, and the slip planes were determined to be (010)
and (00]). Consequently, the dislocations have <xy0> type Burgers vectors.

4.5.1.1 Commercially Pure F ast-Cooled [001]

This commercially pure material contained scattered tangles of dislocations along
with straight dislocations prior to in-situ straining. With straining, the movement of
dislocations was again slow and progressive (Figure 34). The average dislocation
velocity was measured to be ~ 0.1 urn/s. However, in some cases, waviness in the slip
traces was observed in addition to the straight trails (Figure 35 a). Afier a limited amount
of deformation, extensive loops and debris formation were observed (Figure 35 b)
particularly within the slip bands (Figure 35 a).

Figure 36 shows two images of the loop debris that formed over time within the
slip bands in a different sample with similar thermo—mechanical history. As further
dislocations pass through the area shown in figure 36, there was also evidence of some
limited cross-slip away from the {110} planes as indicated by the wavy slip trails. This is
the only condition under which any noticeable degree of cross-slip was observed in the
commercially pure material. Dislocation motion was noticed in the bulk tensile direction
(caused by rotation of the tensile axis) and in other directions as well. Dislocation motion

occurred on both {010} and {110} planes and the line direction was <100>.

89

 

 

Figure 34. Dislocation motion in commercially pure fast-cooled [001] oriented NiAl.
The mobile dislocations are referenced with the pre-existing dislocation lying in the plane
of the foil and seen at the lefi side of the micrographs. Planar slip is evidenced with the
slip-traces left in the wake of dislocations.

90

 

Figure 35. Images of dislocations generated via in-situ straining of air-cooled
commercially pure single crystal oriented to [001] illustrating (a) both wavy and straight
dislocation trails and (b) extensive accumulation of dislocation debris following straining.

91

().51r "1
<——>

0.511 m
H

 

Figure 36. Limited cross-slip was observed in commercially pure fast cooled [001]
oriented NiAl; loops and debris formed over time with straining.

92

4.5.1.2 Commercially Pure Slow-Cooled [001]

Commercially pure specimens cooled slowly within the furnace through the
critical temperature region of embrittlement (673 K to 473 K) exhibited somewhat mixed
dislocation structures prior to deformation. Heavily tangled pre-existing dislocations
were observed in some areas, whereas straight isolated dislocations were pre-dominant in
'others. Again, none of these pre-existing dislocations were seen to move. In most cases
the stress induced mobility of the dislocations ceases when they interact with the tangles.

In some instances the mobile dislocations moved within narrow channels of slip
bands and through the pre-existing tangles with relative ease. Figure 37 (a) shows an
example of dislocations being able to channel through such tangles during straining.
Once the mobile dislocations overcome the interaction with the pre-existing tangles,
newly generated dislocations migrate relatively easily through the pre-existing
dislocations (figure 37 b and c). Dislocation motion with planar slip was noticed in a
different commercially pure slow-cooled material as shown in figure 38, where mobile
dislocations are referenced with a pre-existing dislocation lying in the plane of the foil.

Figure 39 shows two sequential images of the same area of a different slow-
cooled. [001] oriented commercially pure specimen. In this case, dislocations lying in the
plane of the foil have moved as noted by the numbers. Like the other commercially pure
materials, extensive loop debris have developed in the slip bands with strain and the
dislocations were bowed out, generating additional dislocations through the operation of
single-ended Frank-Reed sources (figure 40). While the dislocation motion was again
slow and deliberate in this material, the measured average velocity was slightly higher

than 0.1 urn/s.

93

 

Figure 37. Prior to deformation, heavily tangled dislocations were observed in many
areas of the commercially pure [001] orientated fumace-cooled material. At the onset of
straining, dislocations channeled through the tangles (a). With continued deformation,
newly generated dislocations slipped through the pre-existing dislocations with relative
ease (b) and (c).

94

 

Figure 38. Dislocation motion in commercially pure [001] oriented NiAl. The mobile
dislocations are referenced with the pre-existing dislocation lying in the plane of the foil
and seen at the left side of all the micrographs. Planar slip is evidenced with the slip-
traces left in the wake of dislocations. Most of the mobile dislocations seen in (b) were
not present in (a) and left the field of view in (c).

95

 

Figure 39. Sequence of dislocation motion in commercially pure slow cooled [001]
oriented NiAl. Dislocations (marked with numbers) lying in the plane of the foil move in
a slow and uniform manner.

96

 

Figure 40. Loops and debris formation is evident in slow-cooled commercially
pure [001] oriented NiAl. Dislocation multiplication by operation of single-ended
Frank-Read sources takes place in the substructure.

97

Although a few wavy slip traces were noticed, straight slip traces were
predominant in all commercially pure materials. Loops and debris were also observed
early in the deformation process. In this case, similar to the previously discussed air-
cooled sample, the slip planes of the mobile dislocations appeared to be {100} and
{110}. The line directions of the glissile dislocations were near [100]. No significant
difference in the dislocation behavior of micro-tensile specimens was observed between
the slow and fast cooled specimens, although the bulk mechanical properties differed
significantly for such thermal treatments.
4.5.1.3 Commercially Pure F ast-Cooled [110]

Dislocation movement in this material was very slow with an average velocity
less than 0.1 urn/s. Dislocations appear to be pinned, and slipped by bowing out fi'om the
ends at the foil surface (figure 41 a). Slip was very planar and cross-slip was seldom
observed. This is indicative of the nature of difficulty lying with dislocation movement
in commercially pure material. Figure 41 b shows a dark-field image of the pre-existing
dislocations lying in the plane of the foil. These dislocations again only impeded further
dislocation movement to a limited degree. With continued straining loops and debris
developed in the microstructure as well as dislocation multiplication took place (figures
41 c and (1).

Figure 42 shows a series of images characterizing the initial stages of dislocation
motion. This is clearly evident that the movement rate was sluggish at the onset of
straining and it required an increased deformation for continuous generation and motion
of dislocations. Again, pre-existing dislocations did not significantly inhibit the mobile

dislocations that glided on well-defined {110} planes. A similar situation was observed

98

05pm
«Q—Dr

 

Figure 41. Commercially pure fast-cooled [110] oriented NiAl. (a) Dislocations slip in
the marked direction in slip bands by bowing out from the dislocation ends at the foil
surface. (b) A dark-field image showing a few mobile dislocations (arrowed) and planar
arrays of pre-existing dislocations. (c) Dislocation multiplication is taking place. (d)
Loops and debris are also evident.

99

 

Figure 42. Series of images in commercially pure fast—cooled [110] oriented NiAl
characterizing the initial stages of dislocation motion. Figures (c) and ((1) show
subsequent slip in well-defined narrow slip bands.

100

in another sample (figure 43) where dislocation movement rate was not identical in all
areas of observation. The slip plane was identified as {110} and the line directions of the
mobile dislocations were close to <001>.

4.5.1.4 Commercially Pure Slow-Cooled [110]

As with other samples, application of strain led to dislocations being generated at
the foil edges and in the foil interiors. Figure 44 shows a sequence of dislocation motion
in the slow cooled material. Dislocations marked 1-4 display the motion of the
dislocations. Again the slip was very planar, and in the early stages of deformation the
dislocation motion is restricted to narrow slip bands. No cross-slip has been observed.

Trace analyses of these well-defined slip traces indicate that the slip planes are
{110}. Some of the dislocations, marked in figure 44 b, appear doubled as a
consequence of the portion of dislocations moving during the exposure of the plate. The
small increment of slip, over the length of the dislocations, displayed by these double
images is consistent with overall mobility of dislocations in this material. The average
measured rate of dislocation movement was approximately 1.3 urn/s.

With strain, the dislocation interactions within the slip bands increased and they
became tangled; as a result of which extensive loop debris accumulated. These debris
inhibited further dislocation motion as new dislocations gliding in the bands had to bow
out from local pinning points caused by debris (figure 45). While these debris did
impede further dislocation motion, it could not eliminate it. The bowed out dislocations
were able to break free from the obstruction in many cases. Additionally, many new
dislocations were produced in slip bands through the bowing-out or pinching-off of the

dislocations.

101

0.51m]

 

Figure 43. Dislocation motion in commercially pure fast-cooled [110] oriented NiAl.
Series of images (a) through ((1) reveal dislocation mobility with planar slip and relative
difference in the velocity of the dislocations.

102

 

Figure 44. Sequence of dislocation motion in commercially pure slow-cooled [110]
oriented NiAl. Dislocations marked 1-5 display the relative motion of the mobile
dislocations. Note the very planar slip confined in narrow slip bands. Dislocations
marked in figure (b) appear doubled as a consequence of portions of dislocations moving
during exposure of film.

103

 

Figure 45. Commercially pure slow-cooled [110] oriented NiAl. With increasing strain
dislocations became tangled and loops and debris accumulated. Bowing out of the
dislocations show that the motion was not totally inhibited by the debris.

104

Figure 46 shows an example of this phenomenon in a sequence of images
captured using the video system. Figure 46 a-c show a dislocation loop (arrowed)
expanding under the applied stress. (The contrast of the dynamic image was enhanced
using Adobe PhotoShop.) In figure 46 d, the loop has pinched-off and was set to expand
again, while a neighboring loop has expanded under the applied stress. It is interesting to
note that despite being heat treated in an embrittling fashion, dislocation generation,
motion and multiplication were possible in all of the slow cooled commercially pure
material. The difference in bulk mechanical behavior of the commercial purity material
is not well correlated with the dislocation behaviors in the micro-tensile specimens.

4.5.2 High purity specimens

In general the pre-existing dislocation density was much lower for the high purity
specimens than for the commercially pure specimens as shown in figure 47 a. Prior to in-
situ straining, planar isolated dislocations were observed in different regions of the foil
(figure 47 a and b). In general figure 47 shows presence of a few loops with some
pinched-off dipoles in slow cooled high purity material. With straining, dislocations
began to nucleate from the foil edge and slip into the material; dislocations were also
generated from deep within the foil but the sources could not be observed due to foil
thickness. Both planar slip and cross-slip was observed in high purity materials; cross-
slip is evidenced by waviness of slip-traces left by the mobile dislocations as they change
slip planes [100, 101].

The most significant difference between the observed behavior of the high and
commercial purity materials was the observed dislocation velocities. The observed

velocities (as reported in Table V) in the high purity materials were at least an order of

105

 

Figure 46. Multiplication of dislocations through operation of single-ended Frank-Read
sources in commercially pure slow-cooled [110] oriented NiAl. Images captured
dynamically with a video system showed the arrowed dislocation loop expanding under
the applied stress (a-c). The next image (d) shows that the loop has pinched-off and is set
to expand again while a neighboring loop has also expanded. (Quality of the picture is
regretted and Adobe PhotoShop was used to enhance the contrast of a dynamic image)

106

 

Figure 47. Pre-existing dislocation morphology in high purity slow-cooled NiAl showed
mostly isolated dislocations; a few loops were also seen along with some pinched-off
dipoles. (a) A typical area of [110] oriented single crystal with much lower dislocation
density compared to commercial purity specimens. (b) An area with unusually higher
cencentration of dislocations in a slow-cooled [001] oriented specimen before straining.

107

magnitude greater than in the commercial purity materials and indicate significant
differences in dislocation mobility. Sometimes dislocations were seen to move at a speed
where the critical flicker frequency of human eye was not adequate to capture the
progressive movement of the dislocations. In comparison, for commercially pure
materials, it was possible to track movement of isolated line segments.

As mentioned at the beginning of this section that the pre-existing dislocations
remained sessile throughout the process of in-situ deformation, even after the passage of
cracks physically divided the foil. This may suggest that the pre-existing dislocations are
immobile or pinned. However, this may also just be a consequence of the pre-existing
dislocations being in orientations disadvantageous for slip in the in-situ studies.
4.5.2.1 High Purity Fast-cooled [001]

Motion of dislocations in the high purity NiAl was typically characterized by
clearly evident slip traces left in the wake of dislocations. The dislocations were seen to
move at a variety of rates. Many dislocations were found to move at much higher rates
relative to other observations for commercially pure materials in this study. The very
high velocity of the mobile dislocations could not be determined accurately because the
mobile dislocations could not be captured in two successive frames, but is estimated in
the range of 5 — 25 urn/s.

It may be mentioned that the dislocation velocity was calculated from the TV
screen where images are seen at the rate of thirty-two (32) frames per second. Critical
Flicker/fusion frequency (CFF) for normal human vision is determined to be in the range
of 40 — 90 Hz depending on the intensity of the light [106]. This implies that if any

dislocation moves between two successive frames, human eye may or may not capture it

108

depending on the speed at which the object moves and also the intensity of light at the
screen. Therefore, dislocation velocity was estimated to the closest approximation for
instances where dislocation velocity could not be calculated precisely.

Figure 48 shows both planar slip and development of loop debris at different part
of the same foil. In some cases orthogonal slip was observed for dislocations with line
direction perpendicular to the foil surface (figure 49). Dislocations lying in the plane of
the foil (figure 50) were also seen to move quite rapidly. There must be high, local
residual stress developed due to straining of the thin foil causing movement of such
dislocations that have zero Schmid factor. The slip plane and the line directions of the
mobile dislocations were identified as {112} and <110> respectively.

4.5.2.2 High Purity Slow-cooled [001]

Dislocations were seen to move with ease by rapid jumps. The velocity of the
mobile dislocations was measured in the range 4 — 20 urn/s. Dislocations were generated
in the thicker parts of the foil and the movement of the dislocations was assisted by cross-
slip. The direction of motion (not to be confused with the slip direction) of the mobile
dislocations had changed fiOm [001] to [111]. The observed slip traces in figure 51 show
a dislocation cross slipping past its neighboring dislocations on its way to exiting the foil
at the edge. In some other cases dislocations were also seen to generate at the foil edge
and move towards the thicker part of the foil; it was very difficult to follow the
dislocation activity in those thicker areas.

Some of the samples exhibited a very different response with the applied load and
very slow and deliberate movement of dislocation was observed (figure 52). Increased

deformation resulted in the development of loops and debris in slip bands, along with

109

 

Figure 48. Dislocation motion at the onset of straining in high purity [001] oriented fast-
cooled material. Micrographs were captured at two different parts of the foil when
dislocation generation and movement were noticed from the edge of the foil. While slip
was mostly planar and debris were not readily formed in one part of the foil (a), the other
part shows a significant amount of dipole debris in well-defined slip bands (b).

110

 

Figure 49. Dislocations normal to the plane of the foil moved in orthogonal directions in
high purity [001] oriented fast-cooled material. The contrast fi'om the slip traces of such
end-on dislocations was not as sharp as observed for dislocations lying at some angle to
the plane of the foil. The disturbance created on the oxide layer was less pronounced for
being captured as well-defined slip traces during the exposure time of still photographs.

11]

 

Figure 50. Dislocations that are lying near the plane of the foil (marked 1, 2, 3, and 4) in
figure (a) were found to move quite rapidly in high purity [001] oriented fast-cooled
material and not seen in figure (b).

112

 

Figure 51. High purity slow-cooled [001] oriented NiAl. (a) Dislocations generated
within the foil and moved towards the edge. (b) The arrowed-dislocation cross-slipped
past its neighbors.

113

 

Figure 52. Slow movement rate noticed with dislocations generated in high purity slow-
cooled [001] oriented NiAl. Dislocations moved at such a low rate that movement of
individual segments of dislocations were apparent.

114

activation of single-ended Frank-Read sources. Figure 53 shows a sequence of such loop
and debris formation. Orthogonal slip was also observed (figure 54). The slip planes
were identified as {112} and {110} with [101] line directions of the mobile dislocations.
4.5.2.3 High Purity F ast-cooled [110]

Figure 55 (a) shows an area in the high purity fast cooled NiAl, which possessed
uncharacteristically high numbers of pre-existing dislocations. Most of the pre-existing
dislocations were nearly straight and isolated; a few loops were also seen. In high purity
material dislocations began to move at a very rapid rate with straining and glide past all
of the pre-existing dislocations. It appeared that the pre-existing dislocations could not
inhibit the slip of the mobile dislocations. Mobile dislocations have escaped the field of
view in figure 55 b leaving wavy slip traces.

Similar to [001] oriented fast cooled material, dislocations moved at a very high
rate for [110] orientation also. The mobile dislocation velocity could not be determined
with acceptable uncertainty and approximated in the range of 10 — 25 urn/s. The slip
planes were identified as {001}.
4.5.2.4 High Purity Slow-cooled [110]

Dislocation motion was characterized by clearly evident slip traces left in the
wake of dislocations (figure 56 a). With straining, dislocations moved by rapid jumps
and dense slip bands developed from the foil edges. Dislocations, which are normal to
the foil, moved in slip bands and slipped through the pre-existing dislocations lying in the
plane of the foil (figure 56 b).

Pre—existing dislocations did not seem to inhibit the motion of such end-on mobile

dislocations. The mobile dislocations were found to move on {112} planes and had

115

(1.511131
H

 

Figure 53. (a) Dislocations generated within the foil in high purity slow-cooled [001]
oriented NiAl. (b) Mobile dislocations developed loops and debris as a result of
dislocation motion and interactions.

116

 

Figure 54. Dislocations moving in the foil leaving well-defined trails of planar slip in
high purity slow-cooled [001] oriented NiAl (figures a and b). Line directions of the
mobile dislocations were close to the foil normal. Debris were developed at a slower rate
and orthogonal slip takes place. Figures (c) and (d) reveal orthogonal dislocations lying
in the plane of the foil after in-situ straining. The dislocation density decreased as the foil
edge was approached.

117

 

Figure 55. High purity fast-cooled [110] oriented NiAl. (a) The pre-existing dislocation
morphology shows high number of nearly straight, isolated dislocations with a few loops.
(b) Dislocations moved at a very rapid rate with straining and have slip past all pre-
existing dislocations. Cross-slip of the dislocations is evident fi'om the wavy slip trails.

118

 

Figure 56. Evolution of dislocation slip in high purity [110] oriented slow-cooled NiAl.
(a) Pre-existing dislocations lying in the plane of the foil did not move while newly
generated dislocations (lying normal to the plane of the foil) glided leaving visible slip
trails. (b) Slip bands were more evident after further deformation.

119

<11 l> line directions. In a different sample dislocation motion was observed (figure 57)
in different parts of the foil at the onset of straining. Dislocations that were generated
from the foil-edges moved readily into the foils, ofien moving through or past the pre-
existing dislocations. Figure 57 a shows the dislocations, which were generated at the
nearby foil edge migrating into thicker areas of the foil. Note that the dislocations were
bowed out to a degree in the direction of motion with the ends (at the top and bottom of
the foil) lagging somewhat behind. Figure 57 b shows not only dislocations that had
slipped, but also slip trail, left in the wake of dislocations. These trails are the
consequence of the dislocations disturbing the top and bottom of the foils as they pass.
The slip trails in the present case are wavy, indicating that cross-slip is taking
place, and the dislocations must be screw in nature. Trace analysis of the dislocations
shows the line directions were near <111> and the average slip plane was near {112}.
The dislocations were seen to move at a variety of rates. Many dislocations were
found to move at high rates (relative to other observations in this study) with a measured
velocity of approximately 13—17 urn/s. Other dislocations were found to move at slower
rates, in particular in association with dislocation pile-ups. With continued straining,
significant levels of tangles developed in the slip bands (figure 58). Unlike commercially
pure material, dislocation tangles did not seem to inhibit further dislocation motion. As
the dislocation tangles continued to develop, dislocation multiplication took place within
the slip bands, by operation of single ended Frank-Read sources. With time, extensive

loop debris formed in the bands (figure 59).

120

 

Figure 57. Images of dislocation motion in high purity slow-cooled [110] oriented NiAl.
(a) Dislocations generated at nearby foil edge gliding into thicker areas of the foil. The
direction of motion is marked on the figure (b) that was different than the tensile
deformation axis (TA) marked on figure (a). (b) Wavy slip trails left in the wake of the
dislocations indicate a variation in the slip planes of the slipping dislocations.

121

 

Figure 58. With increasing deformation, dislocation tangles developed in the high purity
[110] oriented slow-cooled NiAl. Unlike commercial purity material dislocation tangles
did not inhibit movement of the mobile dislocations (marked md in figure 58 a) in high
purity specimens.

122

 

Figure 59. With increasing deformation, loops and dipole debris developed in the slip
bands in the high purity [110] oriented slow-cooled NiAl.

123

4.5.3 Disc-Straining of High Purity Specimens

A few selected 3 mm disc samples from high purity single crystals were deformed
in the tensile stage as described in the experimental section (3.7). The straining was
never continued through the failure of the samples and always stopped after a noticeable
extent of dislocation activity. Samples were transferred to a double-tilt holder following
straining. Thus, all three primary zone axes could be utilized with increased flexibility of
tilting to the desired angles; also difficulty of using invisibility criterion to characterize
mobile dislocations could be minimized.

Mobile dislocations in figure 60 were observed in reference to the pre-existing
dislocations shown at the bottom-left comer of the micrograph (figure a). The
dislocations are close to invisibility in figure (b) and visible with weak-contrast in figure
(c) suggesting [001] Burgers vector. Similarly, in a different fast-cooled high purity
single crystals sample, mobile dislocations with [100] Burger’s vector observed in narrow
slip bands in figure 61. Several pre-existing dislocations were seen in the field of
observation on both sides of the slip bands. As an example, dislocations marked ‘1’ and
‘2’ were identified to have [010] Burgers vector.

Figure 62 reveals a complex situation for mobile dislocation characterization in a

different high purity slow-cooled NiAl sample where complete invisibility could not be

achieved utilizing three different zone axes such as [001], [011] and [TTI ]. There may

be two different types of mobile dislocations. Dislocations somewhat confined in slip

bands may have [IOT] Burgers vectors for being in weak contrast in figure (b) and and

complete invisibility in figure (c). Mobile dislocations, to the right of the slip bands, may

have [1 1T] Burgers vectors for being out of contrast in (c) and weak contrast in (d).

124

 

Figure 60. Mobile dislocation characterization in high purity slow-cooled NiAl. The
mobile dislocations are referenced with the pre-existing dislocation lying in the plane of
the foil. Dislocations with good contrast in (a) are observed with weak-contrast in (b)
and become invisible in (c) suggesting [001] Burgers vector.

125

 

Figure 61. Dislocation mobility was observed within a narrow slip band. Mobile
dislocations were visible in good contrast in figures a) and c) and appeared with weak
contrast for figures (b) and (d). These dislocations were characterized to have [100]
Burgers vector. Likewise pre—existing dislocations (marked 1 and 2 on both sides of the
slip-bands in d) are seemed to have [010] Burgers vector for becoming invisible in
figures (a) and (c).

126

 

Figure 62. Mobile dislocations in high purity slow-cooled NiAl. In absence of complete
invisibility, the Burgers vector could not be conclusively determined. Two types of
mobile dislocations were present. The ones somewhat confined in the slip bands have

[101] Burgers vectors from figures (b) and (c) whereas dislocations to the right side of

the slip band may be of [l l 1] type from figure (c) and (d). The line direction of the pre-
existing dislocations lying in the plane of the foil was determined to be [010] from
figures (a) and (c).

127

Unlike in-situ straining of tensile specimens, no cross-slip was observed during disc
straining of high purity specimens.

Although observation of <lll> dislocations is not very likely for <110>
orientation, the same type of dislocations was found to be mobile in other samples too.
Figure 63 reveals dislocation mobility in high purity fast-cooled specimens. Mobile
dislocations were observed in reference with two pre-existing dislocations. Dislocations
were seen to move in a slip band in figures 63 (a) and 63 (c) and those were out of

contrast in figure 63 (b) and observed with residual contrast in figure 63 (d). This

suggests that the mobile dislocations have [1 11] Burger’s vector.

Figure 64 shows a sequence of micrographs to find [100] dislocations in high
purity fast-cooled specimens. The field of view contains unusually high numbers of
mobile dislocations that were moving from the right side of the micrograph to the left.
The slip was planar as seen from the slip traces and no cross-slip has been observed in
this case. All of the mobile dislocations may not be of the same type or there may be
localized stress, which reveal dislocations at different contrast level using the same

diffracting vector. Mobile dislocations in high purity fast-cooled NiAl (figure 65), seen

in contrast in figures 65 (a) and 65 (b), were found to have [101] Burgers vectors for
being in weak contrast in figures 65 (c) and 65 ((1).

Figure 66 shows an area with unusually high dislocation density in fast cooled
high purity single crystal. Dislocations were seen to be mobile in slip bands while some
loops and deformation substructure were also evidenced. Mobile dislocations were
identified as [111] type. In another slow-cooled high purity material [101] dislocations

were found to be mobile as seen in figure 67. Dislocations were found to be mobile in

128

 

Figure 63. Pre-existing dislocations lying in the plane of the foil are used as reference in
high purity fast-cooled single crystal. Dislocations seen in good contrast in (a) and (c)

were observed with weak contrast in (b) and (d) and determined to have [1 1 1] Burgers
vector.

129

 

Figure 64. Mobile dislocation characterization in high purity fast-cooled single crystal.
Dislocations, those were seen in good contrast in (a) and in (b) became invisible in (c)
and were observed with weak-contrast in (d) suggesting [100] Burgers vector.

130

 

Figure 65. The mobile dislocations in high purity slow-cooled single crystals are
referenced with the pre-existing dislocation seen in the right side of the micrographs.
Dislocations seen in good contrast in (a) and in (b) were observed with weak contrast in

(c) and became invisible in (d) implying [101] Burgers vector.

131

 

. ,7.
Figure 66. Images from slow-cooled high pu 'ty NiAl shows mobile dislocations in slip
bands. A significant amount of deformation substructure has been formed and some
dislocation loops are also observed in the substructure. The dislocations within the slip
bands are in extinction in figure (c) and in very weak contrast in figure (b) indicating
mobility of [l l 1] dislocations.

132

 

Figure 67. Dislocations moving in narrow slip bands in slow-cooled high purity MA].
The dislocations are in extinction in figure (c) and in very weak contrast in figure (d)
suggesting mobile dislocations [101] mobile dislocations.

133

narrow slip bands and contrast analyses was performed utilizing the pre—existing
dislocations as well as some stains in the sample. Figure 68 shows an area very much
similar in appearance like figure 66 except for slow-cooled specimens. Mobile
dislocations were identified as [101] type whereas some of the deformation-induced

dislocations and loops were [010] type.

134

 

Figure 68. Images fiom slow-cooled high purity NiAl show mobile dislocations in slip
bands. The mobile dislocations were in extinction in figure (c) and in very weak contrast
in figure (d), suggesting [101] mobile dislocations. Similarly, some of the pre-existing
dislocations or loops may be of [010] type for being in extinction and weak contrast in
figures (b) and (c) respectively.

135

5.0 DISCUSSION

The behavior of the NiAl single crystals studied in the current investigation varied
depending on the impurity concentrations and the imposed cooling rates. The final
cooling rates from elevated temperatures influenced the yield stress significantly in the
commercially pure material, but the influence was not so pronounced in the high purity
material. The differences in mechanical behaviors and the differences in dislocation
morphology following compression deformation of the commercially pure and the high
purity single crystals are explained in the present study in relation to static strain aging.
The dislocation behavior consequent to tensile straining also varied widely depending on
the alloy chemistry, which can also be explained in light of dynamic strain aging.
Important findings from these observations described in the preceding results section will

be discussed in the following subsections.

5.1 Structure-Property Correlations

Dislocation substructures and the response to cooling rate varied significantly
between the deformed and undeformed conditions of the MA] single crystals.
Undeformed specimens exhibited isolated and widely spaced dislocations throughout the
structure; few loops and tangles were observed. The overall dislocation density was
lower for high purity (HP) materials than for commercially pure (CP) materials and
within the same purity single crystals the dislocation densities were similar for both the
fast and the slow-cooled condition in the undeformed state. Reheating at intermediate

temperatures had no apparent effect on this structure, as was shown in figure 16 for the

136

commercially pure material. This observation for undeformed NiAl single crystals
provides evidence that the effect of cooling rate on existing dislocation morphology was
not so pronounced when the dislocation density was low in the substructure. In absence
of any barriers or dislocation tangles, isolated dislocations exhibited some mobility as a
result of stresses generated fi'om beam heating. The slip plane of these mobile
dislocations was identified as {001}, and both <110> and <001> dislocations were
present in the microstructure, with the majority being <001>. The presence of <001>
dislocations has been reported in almost all such previous works except in <001>
oriented single crystals [6 and 7].

High purity and commercially pure NiAl single crystals responded differently to
the compression deformation tests as shown by their yield-stress values reported in Table
III and IV. This difference is most likely due to the difference in the impurity
concentrations of the commercial purity and the high purity materials, as detailed in
Table 11. As such, the solid solution hardening effect (a function of the solute
concentration) of any commercial purity material will always be expected higher than for
any high purity material. hnpurity atoms such as carbon and oxygen occupy the
interstitial sites and induce solid-solution hardening in MA] that is reflected in the higher
yield stress of the commercial purity NiAl than the high purity NiAl single crystals.
Weaver et a1. [87] concluded that carbon and oxygen are the key contributors to the solid
solution strengthening and their contribution is ~ 2500 MPa/at% in commercially pure
NiAl relative to high purity NiAl. The increase in yield stress resulting from alloying
element additions has been extensively dealt by several researchers [6, 7] in the past and

will not be discussed here.

137

As noted in the Literature Review (section 2.2), the mechanical properties of NiAl
at ambient temperature are strongly influenced by impurity concentration, pre-stress,
thermal-history, and the presence of any gettering agent/s. Thermal treatments prior to
deformation of commercial purity single crystals impacted the yield-stress determined in
this present investigation to similar levels as the fracture-toughness properties measured
by Hack et al. [18-20]. For example, the yield-stress was more than doubled by changing
the cooling rate from the fast (1 K/s) to the slow (0.1 K/s) rate following a 1600 K
homogenization anneal, whereas the fiacture-toughness decreases by almost 85% for the
same change. Similarly, for the samples reheated at 673 K, the fracture-toughness
decreases by 65%, with a concurrent 57% increase in the yield-stress when the cooling
condition is changed from fast to slow rate. Figure 12 shows that if the terminal cooling
rate is slow, the ductility decreases and the yield stress increases, regardless of prior
processing parameters for commercial purity specimens. Conversely, even if a sample is
slow-cooled from higher temperature, but then reheated at intermediate temperatures
followed by fast cooling, the fracture toughness increases while the yield stress decreases.

There is not much fi'acture-toughness data available in the literature for high
purity single crystals. No work has been reported replicating experimental conditions in
. line with Hack et a1. [18-20] for high purity NiAl single crystals similar to those used in
the present work. Johnson et al. [107] reported fracture-toughness values in the range of
10-12 MPam”2 for high purity single crystals. But as the influence of prior thermal
treatment on yield properties was not found to be very significant, it may be anticipated
that the fracture-toughness properties are not largely impacted by the thermal treatments

in high-purity NiAl single crystals.

138

In case of the post deformation TEM studies, the deformation substructures, like
the mechanical properties, of both the high and the commercial purity materials,
depended largely on the prior thermal history. Based on the final cooling rate from
elevated temperatures the deformation substructure displayed significant differences.
Heavily tangled dislocations were observed (figures 18 - 20) after deformation of the
slow-cooled commercially pure materials; strong interactions between the pre-existing
dislocations and the newly created dislocations facilitate such tangle formation. The
isolated dislocations, as shown in figure 21, gliding under beam heating conditions,
exhibited limited mobility and ceased movement when they encountered tangled or piled-
up dislocations. The dislocation substructures and the restricted mobility of individual
dislocations imply removal of a good proportion of dislocations from the pool of mobile
dislocations by slow cooling. As a result, during subsequent deformation, larger stresses
were required to propagate newly generated dislocations through the forest of entangled
and pinned dislocations, giving rise to a higher yield stress and limiting the toughening
mechanisms.

Post deformation TEM examination of the fast cooled commercial purity material
did not reveal any dislocation tangles (figures 22 — 24) in comparison to the slow cooled
material. The pre-existing dislocations were not subjected to pinning in this case, as the
cooling rate was sufficient to retain the solutes in solution. As a consequence, during
subsequent deformation, freshly generated dislocations encountered fewer obstacles,
resulting in lower yield stresses and higher fracture toughnesses. Also, it was observed
that the motion of the isolated dislocations by beam heating was not hindered nor

restricted in the fast cooled commercially pure material.

139

In contrast to the commercially pure specimens, mobility of dislocations did not
vary between the fast and the slow cooled high purity samples. In both cases <001> type
dislocations were found to be mobile as a result of beam heating. In most cases isolated
dislocations were observed even after deformation, and the substructures contained few
tangles and very poorly defined dislocation cells. Similar to the commercially pure
material, some <110> dislocations were also observed in the high purity material, as
shown in figure 27 for fast cooled specimens. Dislocation analyses of the slow-cooled
high purity material revealed <111> as the mobile dislocations (figure 28) in addition to
<001> and <110> dislocations. There is no mention in the literature of <111>
dislocations in MA] other than in <001> oriented single crystals. The mobility of these

<111> dislocations will be discussed in the in-situ deformation studies section.

5.2 Static Strain Aging

The observed differences in dislocation morphology and mechanical behaviors
and their correlation are best explained in relation to strain aging. Strain aging is a time
. dependent strengthening phenomenon where a plastically deformed metal or alloy
hardens as a result of elastic interaction of solute atoms with the strain fields of
dislocations. Thus, any alloy system in which interstitial or substitutional atoms are able
to pin the active dislocations and their sources is susceptible to strain aging; low carbon
steel is the most studied and ideal example of this phenomenon [109]. Depending on the
way (prior to or during plastic deformation) the solute atom atmosphere forms around the
dislocations, strain aging may be classified into two distinct types, static and dynamic.

Dynamic strain aging, which becomes more significant when diffusing solUte atoms

140

interact with mobile dislocations, will be discussed at greater length in section 5.4 with
in-situ deformation results.

Static strain aging (SSA) is associated with the return of a sharp yield-point,
increase in yield-stress, and decrease in ductility. Generally, pre-straining, unloading,
aging, and reloading at the same rate of the pre-strain precede SSA. Several theories
explaining the mechanisms of SSA in BCC metals and alloys have been discussed at
length in the literature [108, 109]; they will be summarized here in view of the
similarities between the BCC crystal structure and the B2 system.

There are four different processes that describe SSA, namely, the Snoek effect,
formation of the Cottrell atmosphere, Suzuki locking, and precipitate formation. Suzuki '
locking and precipitate formation are not pertinent to the discussion of the present
investigation. Suzuki locking originates from the chemical interaction between stacking
faults and solute atoms. Suzuki locking is prevalent in metals and alloys with wide
stacking fault widths and low stacking fault energies, such as FCC or HCP systems. It is
unlikely that Suzuki locking is a possible mechanism of SSA in NiAl, which has no
evidence 'of any stacking faults and is a high stacking fault energy system. Precipitate
formation is expected in any alloy system supersaturated with solute atoms. Precipitation
on dislocations during aging results in pinning and gives rise to strengthening, although
overaging causes coalescence of the precipitates to take place by Oswald ripening and
induces softening. Since no gettering elements are used and no precipitates of any
second phase have been observed in the present investigation, precipitation can be ruled

out as the mechanism of SSA in this investigation.

14]

The Snoek effect involves the strain-induced ordering of interstitial atoms around
dislocations when the interstitial atoms (typically occupying the octahedral sites at the
center of the cube edge or cube face) jump from one position to another in the lattice
[110-111]. Application of any external stress along one direction of the unit cell results
in enlargement of the interstitial sites along that axis and an associated decrease in the
size of the openings perpendicular to that particular axis of elongation. Thus, the atoms,
which have no favorable site occupancy in an otherwise strain-free lattice, are forced to
move/jump to new positions in the slightly deformed and tetragonal lattice. This type of
strain induced ordering gives rise to interaction between dislocations and solute atoms,
resulting in an increase in yield stress.

The Cottrell atmosphere [112] is a solute environment around dislocations that
forms during static strain aging and grows depending on the time and temperature of the
aging process. The solute atmosphere is formed by elastic interaction between interstitial
atoms and the strain fields of dislocations. In fact, this elastic interaction is the driving
force for diffusion of solute atoms to the strain-fields of dislocations and results in
decreasing total energy of the system. Thus, the dislocations are locked by solute atoms -
requiring additional stress to move them. Cottrell atmosphere formation increases the
yield stress by a higher magnitude than the Snoek effect [113].

Impurities, such as carbon and silicon, play an important role in controlling the
strain-aging response of Ni-Al alloys. It was observed by Weaver et a1. [85] that the
strain aging effect of carbon is more pronounced in presence of silicon than without it.
Weaver [80] carried out a systematic study of several different polycrystalline NiAl

alloys to find out the species responsible for strain aging. This study found carbon to be

142

the primary element responsible for strain aging rather than nitrogen or oxygen. This
strain aging response was illustrated by development of yield-point phenomenon as
carbon doping was increased in high purity NiAl alloys. Noebe and Garg [114] made a
similar observation using powder processed NiAl. They found no sharp change in the
yielding pattern in the stress-strain curves of a nitrogen-doped alloy (904 ppm N, 347
ppm 0 and 57 ppm C) as opposed to a sharp change in the yield stress and load drop for a
carbon-enriched alloy (143 ppm C, 6 ppm N and 227 ppm 0). Interestingly, high purity
alloys aged for longer times exhibited serrations in the stress-strain curves [81, 83],
which were presumably a result of the lower concentration of interstitials requiring longer
diffusion distances to effect pinning of dislocations.

The commercial purity material used in this present study is also enriched with
carbon (Table II), and in the presence of silicon a strong strain aging response is expected
[85]. The presence of silicon could not be avoided in the commercial purity materials as
silicon gets picked up by the growing single crystals fi'om the alumina-silicate shell
molds in the Bridgeman process. The increase in yield stress associated with slow
cooling is consistent With all earlier findings, including those of Hack et a1. [18-20].
Presumably the mechanisms responsible for strain aging are both the Snoek effect and the
Cottrell atmosphere, as the other two SSA mechanisms are ruled out in the present case.
As the Snoek effect is not associated with long-range diffirsion and gets completed within
the time interval of one atomic jump of the species responsible for pinning, it is expected
that the Snoek effect precedes the Cottrell atmosphere formation. It has been reported
that two plateaus are observed in a plot of increase in yield stress as a function of aging

time [113]. These plateaus are consistent with a combination of Snoek effect and Cottrell

143

atmosphere formation. Weaver [80] observed that the time required for one interstitial
atom to make one jump corresponds to the first inflection point between the two stages of
the above-mentioned plot.

In contrast to slow cooling, fast cooling trapped the interstitial elements in
solution and the solute atoms did not have ample time to come out of the solution to
interact with the dislocations. That is, there was very little Cottrell atmosphere formation
during fast cooling resulting in lower yield stresses in these samples. This observation is
also in conformance with the findings of Hack et al. [18-20] who found that it is possible
to minimize the effect of strain aging by faster cooling. It should be noted that if the
cooling rate is so rapid that vacancies are entrapped, the yield stress could increase,
resulting in additional resistance to dislocation motion [3 8]. It is important to realize that
"the vacancy hardening effect acts in addition to the solid solution hardening effect.
However, the fast cooling rate applied in the present study is much lower than the cooling
rate of quenching necessary for vacancy entrapment. In some cases a few faceted voids
are observed in the substructure [33]. These voids are more common in the quenched
materials. During subsequent reheating of the quenched specimens, entrapped
supersaturated vacancies can coalescence with each other to form these voids and result
in decreasing yield stress and lower work hardening rates [7]. (Examples of such void
formation in the present study are shown infigures 62, 64 and 69.) As the cooling rate
increases, the degree with which solute atoms interact with the dislocations decreases,-
which most likely is the cause of the decrease in yield stress for faster cooled materials.

As mentioned earlier, the response of the high purity material to static strain aging

was quite different from the commercially pure material. This difference in response, as

144

(3.0211 m

 

Figure 69. Voids observed in the substructure of high purity [110] oriented material
reheated to 673 K following homogenization anneal of 1600 K. Three different samples
of the same material revealing isolated and cluster of voids at different magnifications.

145

measured by the mechanical behavior, was reflected in the resulting dislocation
substructures. The high purity material was less susceptible to strain aging primarily due
to less intense interaction of the solute atoms with the pre-existing dislocations that stems
from a lack of ample solute atoms to provide pinning. This weak interaction is also the
root cause for the lower yield stress in high purity materials. As the concentration of
solute atoms was low in the high purity material, more time and longer diffusion
distances were required to pin the existing dislocations to cause strain aging.

It has been discussed that solute segregation to dislocation cores is one of the
mechanisms for strain aging. For strain aging to take place, the local solute concentration
‘c’ at the core of the dislocation has to be higher than the bulk solute concentration ‘co’ of

the crystal. and the two concentrations are related by the following equation [115]:

 

c=coexp(-¥Ifl“) (l)

where Wm is the dislocation solute binding energy, R is the universal gas constant, and T
is the temperature. Therefore, for a constant temperature, the solute enrichment at the
core of a dislocation is directly proportional to the bulk solute concentration or CO.
According to F ick’s Law of Diffusion, the mass flux is proportional to the concentration
gradient. Thus, the probability of pinning a mobile dislocation by solute atoms is higher
in the commercially pure material than in the high purity material.

Some effect of strain aging is still observed in the high purity material as
evidenced by the fact that when these specimens were heated to an intermediate
temperature of 673 K and slow cooled, the yield stress increased in comparison to the
fast cooled samples (Table IV). In such situations the slower cooling rate provided

sufficient time for the diffusing atoms, particularly carbon, to travel longer distances and

146

lock some of the pre-existing dislocations. However, the availability of solute atoms is
inadequate to pin a good proportion of pre-existing dislocations in the high purity
material at the cooling rate provided by the slow cooling used in the current investigation.

Thus, the solute atom concentration and the cooling rate complement each other
to provide the static strain aging effect. The underlying mechanism of strengthening is
the same in both purity alloys; only the degree of hardening differed due to the lesser
availability of impurity atoms to interact with the pre-existing dislocations in the high
purity material. Hence, the difference in mechanical behavior between the commercial
and high purity materials stems directly flom the difference in the concentration of the
impurities.

In a related study on dislocation structures of NiAl single crystals, B-C. Ng [104]
found that the influence of cooling rate on mechanical properties could also be explained
with the help of electron channeling contrast imaging (ECCI). He observed fewer
dislocations ahead of crack tips in slow-cooled materials and suggested that solute
pinning made it difficult to generate or emit flesh dislocations at the crack tip.
Generation of fewer dislocations at the crack-tip resulted in a smaller plastic zone, and
effective crack-tip blunting could not occur. Instead, cracks continued to propagate in a
brittle manner, resulting in failure. In contrast to slow cooled commercially pure NiAl,
Ng et al. [116] have observed large numbers of dislocations ahead of crack-tips in fast
cooled samples, resulting in more ductility in the fast cooled material. For high purity
NiAl, ready emission of dislocations flom crack-tips was observed for both slow and fast
cooling rates [117]. Thus, size of the plastic zone was not a fimction of the cooling rate

for high purity NiAl single crystals as opposed to commercially pure material.

147

All of these observations lead to an understanding of the higher ductility in the
high purity NiAl single crystals. Levit et al. [26] reported elongations as high as 25% in
high purity materials. This extensive straining was facilitated by predominantly <001>
dislocations. However, in the present study, which used similar purity material, <111>
dislocations are observed in addition to <001> and <110> dislocations. This may make
the slip process easier, which in turn may increase the ductility in high purity materials.
Strain aging was not so pronounced in the high purity material, even after reheating to
intermediate temperatures. Although in this study the material was not deformed to
failure and the tensile ductility was not determined, considering the source of the single
crystals and their impurity concentrations, the high purity NiAl alloys would be expected

to have a higher ductility than the commercially pure NiAl alloys.

5.3 In-situ Straining TEM and Dislocation Behavior

As a consequence of tensile strain, new dislocations were generated both at the
foil edges and in the foil interiors (in areas too thick to image) in the specimens. These
fleshly generated dislocations moved irrespective of thermal treatment, orientation, and
alloy purity in all specimens tested in-situ in the current investigation. Even though NiAl
is plagued with low room temperature ductility, dislocation multiplication took place in
all cases. Significant differences in the dislocation velocities and manner of movement
were observed between the high and the commercial purity materials. The observed
differences in dislocation behaviors were presented in the results section (4.5). Some of
these differences in the dislocation behaviors of the materials were heavily dependent on

the alloy purity and will be discussed at a greater length.

148

Prior to deformation, the high purity (HP) specimens contained a lower
dislocation density than the commercially pure (CP) specimens. Planar arrays of pre-
existing dislocations, like those shown in figures 16 and 47, were observed in many parts
of the foils and remained sessile throughout the straining process even after the passage
of cracks physically divided the foil (figure 31). These dislocations, lying in the plane of
the foil, were sessile due to their orientation with respect to the load-axis and plane of the
foil, which implies that the pre-existing dislocations, being in orientations
disadvantageous for slip in the in-situ studies, had resolved shear stresses equal or close
to zero and could not move. Sometimes dislocations lying in the plane of the foil (such
as figure 50) moved very rapidly in the high purity material. Such observations are
considered to be special cases. There are two conditions that may have resulted in
movement of such dislocations with apparent zero Schmid factor (if their glide plane is
the same as the foil or parallel to the loading axis). First, the rotation of the tensile axis
during straining, or second, high localized residual stresses developed around the
dislocations. as a result of tensile deformation of the thin-foil specimens. Regardless,
movement of dislocations lying in the plane of the foil is considered an exception, and in
general such dislocations did not show any activity/mobility throughout the in-situ
straining process.

Because of the thin foil sample geometry, the magnitude of the localized residual
stress in the thin-foil was not constant. As a result, the resolved shear stress on all the
mobile dislocations was not uniform and dislocation movement rate was different at
different parts of the foil. Figure 51 shows such an example, where one of the mobile

dislocations cross—slipped past its neighbors on its way to the foil edge. Other

149

 

 

dislocations generated flom similar sources took longer times to exit the field of view as
well as the sample edge. Non-uniform distributions of the pre-existing dislocations and
variation of the foil-thickness at different areas are thought to be responsible for such
behaviors. In other words, in a typical thin foil all the edges of the foil are not equally
effective in generating dislocations. Thus, due to the complex state of stress in the thin
foils, the obServed dislocation motion unfortunately could not be related to the
macroscopic stress and strain.

The most significant observation in the high purity material was the velocity of
the dislocations at which they moved during in-situ straining. In fact, the rate was always
an order of magnitude higher in the high purity material than the commercial purity
material. In the high purity material, dislocations primarily moved by rapid jumps,
regardless of whether the movement initiated flom the foil-edge or flom the inner part of
the foil. Cross-slip was evident in almost all cases of dislocation motion in the high
purity NiAl, making it easier for the deformation process to take place by gliding screw
dislocations. In some cases the rapid jumps of the dislocations were nearly instantaneous
with the application of strain that following the motion of individual dislocations was
difficult.

Sometimes dislocations were seen to move with a burst phenomenon in the high
puriy material (figure 55), with the mobile dislocations escaping the field of view and
gliding through all pre-existing dislocations lying in the plane of the foil. This illustrates
that pre-existing dislocations were not able to inhibit the slip of mobile dislocations to
any significant degree. In all such cases cross-slip and movement of dislocations took

place _in less than a second. As the probability of cross-slip is proportional to the

150

dislocation length in screw orientations, it may be a fair approximation that all of the '

rapidly moving dislocations were screw-type, and the activation energy of such cross-slip
was very low.

With continued straining, significant levels of tangles developed in the slip bands
in the high purity material. Unlike in the commercial purity material, these tangles did
not inhibit movement of the mobile dislocations (figure 58). Thus, the entangled
dislocations that formed in the high purity material may not be totally pinned. However,
in contrast to the observation of Nagpal et al. [101], no well-defined cell structure
development was observed in the current study. The reason may be that continued
mobility and rapid cross-slip of dislocations through the pile-ups prevented formation of
dislocation cells. Also, as the dislocation tangles continued to develop, dislocation
multiplication by operation of single ended Frank-Reed sources was observed in the slip
bands. Thus, dislocation mobility never ceased in the high purity material with straining.

Although the commercial purity material exhibited sharp differences in the bulk
mechanical properties based on the final cooling rate, no significant difference was
observed in the dislocation behavior during in-situ straining. It was discussed that the
'morphology and substructure of the pre-existing dislocations influenced the mobility of
the newly generated dislocations. Heavily tangled pre-existing dislocations observed in
many parts of the foil did not show any activity during the tensile defamation. These
tangles are the product of previously explained static strain aging and hindered the
dislocation mobility in the commercially pure materials. In general, dislocations
associated with pile-ups were found to move at slower rates; mobile dislocations

generated either flom the foil edge or from the foil interior often ceased to move when

151

I
E'T‘J '14—

 

interacting with these dislocations tangles. While the pre-existing dislocation tangles
inhibited the movement of mobile dislocations in commercially pure materials, they did
not stop it completely. One such example is shown in figure 37 for a slow-cooled [001]
oriented commercially pure specimen. Heavily tangled pre-existing dislocations were
observed in many parts of the foil prior to straining. With straining, the mobile
dislocations, which were confined in narrow slip bands, were able to channel through
such tangles. Once the slip band was established through the tangles, newly generated
dislocations migrated with relative ease through the pre-existing dislocations. Similarly,
dislocation multiplication was observed in slow-cooled commercially pure NiAl alloys
(figUre 40) by operation of single-ended Frank-Reed sources. Therefore, the brittleness
exhibited by the commercially pure alloys may not be related to the inability to generate
additional dislocations through multiplication.

Slip in the commercial purity material was mostly planar and confined to narrow
slip bands. There was some build-up of loop debris in the slip bands after extensive
straining. Cross-slip was observed in limited cases. With continued straining it was
sometimes found that the deformed Specimens developed cracks flom the foil edge,
propagating between slip bands (figure 33 d). This sort of crack generation between slip
bands is a specific example; in general cracks generated in slip bands with continued
deformation. The planar slip and eventual cracking are examples of limited plasticity in
the commercially pure material, where the dislocations were always seen to move at a
fairly constant, slow rate and in a wiggling manner, rather than through quick jumps
observed in high purity materials. The slow movement of the dislocations in commercial

purity material is consistent with the solute drag of the dynamic strain aging effect.

152

Sometimes, dislocations appeared to be bowed out flom the ends at the foil
surface (see figure 41). The long line lengths of the dislocations are divided into several
segments bowed out to a degree in the direction of motion. The radius of curvature of
these bowed out segments, R is inversely proportional to the applied shear stress, I. The
pinning is associated with the surface drag-stress or traction produced by the fiictional
resistance of the oxide layer, which in turn influences the line tension, t, of the
dislocation. For a bowed-out segment this line tension creates a restoring force, which
acts on the dislocation to straighten the length. The line tension can be approximated
[118] as:

=— m

where G is the shear modulus and has a value of approximately 70 GPa for NiAl single
. crystals [6, 7] and b is the magnitude of the Burgers vector. Based on a lattice parameter
of 0.2887 nm for NiAl, the line tension of a [110] dislocation is 2.9 nano Pa. Vailhe et al.
[119] computed the detailed shapes of such curved/arched dislocations and calculated the
Ideal shear stress in NiAl by using the following relationship between line tension, local

radius of curvature and shear stress:

R = (3)

L
br
Combining equations 9 and 10, line tension independent local shear stress can be

obtained:

r-r
11
Sale?

(4)

In the present study, the radii of curvature of the bowed-out dislocations were

approximated between 0.15 to 0.2 pm. It should be noted that the bowed out segments,

153

which are in the form of ogival arches, are more elliptical than circular. In order to
measure the radius of curvatures, the dislocations were conceived to bow out flom the
pinning points and the distances between the pinning points are measured as the diameter
of the curvature. The distance was measured in two ways. First, calibrating a known
distance on the micrograph such as the magnification marker using the Scion Program
[120]. Second, the curvatures were fitted to any known geometric shape such as a circle
or an ellipse. Then the included angle, 0, flom the two pinning points was measured at
the center as shown in figure 70. The diameter of curvature, l, was calculated by the
following equation:

l= 2r sin(0/2) (5)
where, r is the radius of the fitted circle or the average of the major and minor axis if the
shape 'of the curve is elliptical. The local shear stress of such bowed [110] dislocations
were calculated using equation 4 and found to be 67 MPa and 50 MP3 for l-va‘lues 0.15
pm and 0.2 pm respectively. Equation 3 also implies that the smaller the applied shear .
stress, the larger the radius of curvature of the dislocations with other factors remaining
constant. The calculated values of shear stress fall close to the critical resolved shear-
stress values (45 — 60 MPa) measured by Vailhe et al. [119] in MA] single crystals.
These values are in close conformance with the macroscopic critical resolved shear stress
(63.6 MPa for {110}<111> slip system and 67.5 MPa for {110}<lll> slip system) for
slow cooled hard oriented HP single crystals. It may be noted that the macroscopic

critical resolved shear stress values were calculated using the Schmidt’s Law [118]:

P
=——co}t - 6
A scos<|> ()

154

 

 

 

Figure 70. An illustration of measuring the radius of curvature for bowed-out
dislocations in figure 53 (a) using the Scion program [120].

155

 

where P is the applied load acting on a cross section of area A and A and (I) are the angles
of the slip plane and slip direction with the tensile axis.

Dipole formation occurred in NiAl alloys by interaction of cross slipping screw
dislocations on different planes. Vailhe et al. [119] described such formation at
successive steps. According to their model, when a gliding dislocation interacts with
small obstacles, the dislocation is divided into two parts and starts gliding on two parallel
slip planes. Thus, a dipole is formed which closes up when the two parallel slip surfaces
interact with each other to release the dislocation.

The presence of loop debris impeded further dislocation movement, but many
new dislocations were generated when the dislocations broke flee flom the obstacles
(figure 45). Figure 53 b shows several loops that have been left behind by the mobile
dislocations at their point of interaction, as shown in figure 53 a. Figure 71 shows such
loops at different stages of formation in slow cooled high purity [001] oriented NiAl.
Figure 71 a shows how two curved dislocations separated by a jog starts forming a loop,
while figure 71 b shows pinched-off loops and formation of others in progression as a
result of interaction between bowed out segments of dislocations. Shi et al. [121]
proposed a model for dipole and loop formation in NiAl that is schematically shown in
figure 72. This model is consistent with the current study.

The dislocation multiplication process and loop and debri formation in the high
purity material can be explained based on the model of Koehler [122]. According to the
model, screw dislocations cross-slip flom the primary glide plane to a second plane, but
eventually cross-slip back to a plane that is parallel to the primary plane. Depending on

the excursiondistance between the primary glide planes, trailing dipoles or pinning

156

 

 

 

Figure 71. Loop formation at its different stages in high purity slow-cooled [001]
oriented NiAl. Figure (a) shows dislocations separated by jogs are coming close together
to start forming loops and figure (b) shows progression and finally pinched off loops.

157

 

‘\

(c) Pinching-off of a loop

Figure 72. Schematic view of loop and dipole formation after Shi et al. [121].

158

points result, which can act as Frank-Read source [123]. Koehler also described that
multiple cross glide is possible as a consequence of collisions of moving screw
dislocations [122]. In such a situation, the velocity vector, v, and the jogs created on the
secondary plane both lie perpendiculars to the Burgers vector, b, of the dislocation.
These jogs are essentially edge dislocations, which cannot move out of the secondary
glide plane. The edge dislocations can form a dipole depending on the separation
distance between the primary glide planes and the applied stress. Therefore, dislocation
multiplication by cross-glide and dipole formation are not mutually exclusive events,
instead can occur simultaneously.

The dislocation behavior observed during in-situ straining electron microscopy
was found to be independent of the tensile axis of the specimens. The soft oriented
commercially pure material behaved similar to the hard oriented material during in-situ
straining. Like the fast cooled hard oriented samples, limited cross-slip was observed in
the slow cooled materials. Also, similar to the slow cooled hard oriented commercially
pure material, dislocation multiplication was evident in the soft oriented slow cooled CP
material (figure 45). The same situation was observed for the high purity material. The
observed dislocation velocities did not differ significantly with tensile axes of the
specimens. As mentioned earlier, thin foil conditions and localized rotation of the tensile

axes made it difficult to correlate the dislocation behavior with the bulk tensile axis.

5.4 Dynamic Strain Aging

The influence of cooling rate in altering the mechanical behavior and dislocation

characteristics of the high and the commercial purity NiAl single crystals was explained

159

before in light of static strain aging. In this section, the dislocation behavior observed
during in-situ straining will be assessed in relation to dynamic strain aging.

Dynamic strain aging (DSA) is associated with the active formation and migration
of the solute environment around mobile dislocations during plastic deformation.
Serrations, jerkiness, or other discontinuities found in the stress-strain curves determined
in the constant elongation rate compression or tensile tests are indicative of DSA. An
initial low mobile dislocation density and a low sensitivity of dislocation velocity to
stress are considered to be the two major contributing causes for the appearance of
discontinuity in the stress-strain curve [124, 125]. It has been proposed [126, 127] for

materials susceptible to dynamic strain aging that the dislocation velocity (V) is related to

the applied resolved shear stress (1:) as:

V = ID” (7)
where D is the diffusivity of the solute and m is the strain hardening exponent. As D and
m are constants within reasonable limits, dislocations with low resolved shear stress

would exhibit limited mobility. Depending on the strain rate, DSA occurs over a wide
temperature range. For constant rate tensile or compression tests, the strain rate (6‘)

consists of two components, elastic (6‘. ) and plastic (6,, ). Weaver [80] described that for‘

a load drop to occur, the plastic strain rate has to exceed the imposed strain rate. It was
assumed that the plastic strain rate is governed by the Orwan equation for gliding

dislocations:

op: pme (8)

160

 

where pm is the mobile dislocation density, b is the Burgers vector of the dislocation, and
V is the dislocation velocity. A sudden increase in pm, V, or both can result in serrations
or a decrease in stress in the stress-strain curves. Repeated locking and unlocking of
dislocations will increase pm instantaneously. If the solute atoms repetitively diffuse to
dislocations to form dynamic solute atmospheres, serrations occur on stress-strain curve.

Thus, Weaver [80] concluded that dynamic strain aging would become enhanced
when the drift velocity of the solute atoms in the stress fields of dislocations equals V. In
other words, at low temperatures and high strain rates, dynamic strain aging becomes
insignificant as the dislocation velocity largely exceeds the solute velocity; at elevated
temperatures and low strain rates, solute migration rate either exceeds or keep pace with
the dislocation velocity, resulting in negligible dynamic strain aging or no serrations.
Dynamic strain aging can occur between these extremities.

Like SSA, there are several theories explaining the mechanism of DSA. The
Cottrell model [128, 129] predicts that serrated flow takes place when mobile
dislocations, moving at less than a critical velocity, drag along the solute environment
around them. The critical dislocation velocity, Vc, above which dislocations break flee

flom the solute atmosphere, is given as:
Vc = 4 _ (9)

where D is the diffusivity of the solute and l is the effective radius of the solute
atmosphere, which is a function of the solute concentration. The diffusivity, D, is given

by:

D = Do exp(%) (10)

161

 

 

where D0 is the pre-exponential coefficient or flequency factor, Q is the activation energy
(varying between 66 to 115 kJ/mol [85, 86] for carbon diffusion in NiAl alloys), and R
and T carry their usual meanings.

In order to make comparison between the effective radii of the solute atmospheres
in HP and CP alloys, it is important to have an estimate of D for carbon diffusion in NiAl
alloys at ambient temperature. Diffusivity was not measured in the current investigation.
D was estimated to be 3X10'22 cmZ/s at 300 K by extrapolating the experimental values
determined by Weaver [80] and using Microsoft Excel (figure 73). No directly measured
Do data was found in the literature and Do was calculated to be leO’8 cmZ/s; the
activation energy derived flom figure 73 has a value of 78 kJ/mol for C diffusion in NiAl
alloys. In comparison, the activation energy for C diffusion in Fe lattice varies between
84 and 142 kJ/mol [130, 131]. Although the activation energy values for interstitial C
diffusion in Fe and NiAl lattice are similar, the Do determined flom Weaver’s [80] result
does not match the order of magnitude for C diffusion in Fe lattice, which has a value
between 0.02 and 2.2 cmZ/s [130, 131]. In general, substitutional elements are expected
to diffuse slower than the interstitial elements in any lattice by virtue of the higher
activation energy associated with substitutional diffusion. For example, a substitutional
element like Ni has Do value of 077-4 cmZ/s for diffusion in Fe lattice with activation
energy in the range of 240 to 280 kJ/mol [132]. Likewise, ”M has a Do value of 4.5
cmz/s for diffusion in stoichiometric NiAl alloys with activation energy of 307 kJ/mol
[90]. Thus Do, like activation energy, is expected to have similar values for interstitial C
diffusion in Fe and NiAl lattices. Therefore, the D value for C diffusion in NiAl alloys,

based on the Do data derived flom Weaver’s [80] result, is off by few orders of magnitude

162

 

 

 

Measurement of Activation Energy and Frequency Factor
from the experimental values of Weaver [80]

 

-40

LnD

 

y = -9438.5x - 18.081
Q = 78,471 J/mole and Do = 1*10"-8 cm"2lsec

 

 

 

-55 . .
0.001 0.0015 0.002 0.0025 0.003 0.0035
1/T

Temperature (K) D Experimental (cm/‘2/sec) Do (cmA2/sec) Q (J/mol) R (J/mol/K)

300 3522 “3‘03 78471 8314
522 , 2E16 lE-08 78471 8.314
616 ~ 313-15 1E-08 78471 8.314
700 ..ZEIQ 1E-O8 78471 8.314
f f, ‘ _J.‘.‘.1;_;T.3Measured
Estimated

 

 

 

 

Figure 73. Plot of diffusivity with temperature after Weaver [80] and extrapolation of
elevated temperature diffusivity values to 300 K. The flequency factor, Do and the
activation energy, Q were calculated flom the equation of the straight line obtained using

Microsoft Excel.

163

and the D0 value of 1.4X10'8 cmz/s is not consistent with what would be expected for
interstitial carbon diffusion in NiAl alloys. Weaver also agreed on such inconsistency in
private communication. In spite of the problems in determining the D value, the effective
radii, 1 may be qualitatively compared in the two alloys based on the relationship of
equation (9).

According to equation (9), the dislocation velocity is inversely proportional to the
effective radius of the solute atmosphere. This equation is very important in
understanding the difference in dislocation velocities with alloy purity. Above a
threshold of Vc, the drag stress on the dislocations decreases with an increasing
dislocation velocity. As the high purity alloy has a lower global concentration of solute
atoms than the commercial purity alloy, the local solute atom concentration at
dislocations in the HP alloy will be lower compared to the dislocations in the CP alloy.
Therefore, the effective radius of the dynamic solute atmosphere interacting with the
dislocations is also smaller in the HP material compared to the CP material. Based on the
average measured dislocation velocity for the high and the commercial purity NiAl, there
exists an order of magnitude difference between the effective radii of the solute
atmospheres.

The observed dislocation motion was not continuous in the present study.
McCormick [133, 134] developed a model for DSA based on the assumption of
discontinuous dislocation motion and proposed that the interaction of solute atoms with
mobile dislocations takes place when the dislocations wait at obstacles. McCormick’s

model equates the dislocation velocity, v, to the average obstacle spacing, L:

(11)

 

164

 

where tW is the mean waiting time at an obstacle, and tr is the mean flight time between
obstacles. The dislocation velocity can be simplified in cases when the time of flight, tr,

is negligible and v is expressed as:

(12)

ort =

W

V = _L_ _L_
tW v

With this kind of linear relationship between obstacle spacing and waiting time, it
is evident that tw is higher for the commercial purity material than for the high purity
material. For example, with a dislocation velocity of 10 urn/sec and average obstacle
spacing of 0.25 um, tw becomes 25 milliseconds for the high purity alloys compared to
500 milliseconds for the commercial purity alloys with the same obstacle spacing (figure
45) but a velocity of 0.5 urn/sec. This higher value of tw in the commercial purity alloys
provides more time for the solute atoms to pin the mobile dislocations. The higher
probability of pinning mobile dislocations is consistent with the sluggish movement of
dislocations in the commercial purity alloys.

According to McCormick, yield serrations will occur when ta, the aging time

becomes equal or close to tw. Based on the available models [112, 115], McCormick

described ta as the time required to lock mobile dislocations:

c 2” 1th2
to = ' (13)
01C0 3DUm

where C; is the critical solute concentration to lock a dislocation, Co is the bulk solute

 

 

concentration of the alloy, Um is the binding energy as defined by Wm in equation (1), D
is the diffusivity of the solute, and a is a constant with a value of 3. For the
commercially pure material with a higher C0, ta will be lower if everything else remains

unaltered, resulting in enhanced pinning compared to the high purity material. The in-

165

 

situ tensile straining process in the current investigation was intermittent. There was
always a pause or momentary stop after each increment of straining to watch and record
the dislocation activity. During this time stress relaxation took place and mobile
dislocations waiting at any obstacle could be locked. When the aging time exceeds the
waiting time (t8| > tw), dislocations waiting at obstacles will not be locked, causing flee
flow, which is the case for the high purity alloys. Also, it takes a longer aging time to
lock the mobile dislocations when the temperature of the process limits the diffusivity of

the solutes. The availability of more solute atoms provided a decreased t. in the

commercially pure alloys and consequently DSA resulted in restricted mobility of
dislocations.

Manifestation of DSA in the high purity alloys was not evident in the dislocation
behavior. As discussed earlier, long-range diffusion is required for dynamic strain aging
to be effective in the high purity material. One of the conditions for long-range diffusion
to take place is enhanced thermal activation. However, the thermal activation for the
DSA process was controlled by the thermal energy of the electron beam heating and
strain aging could not take place in a dynamic manner with low solute concentration and
slightly elevated temperature. The microscope was always operated at 200 kV and the
physical characteristics, including the thickness of the samples, were similar for both
purity samples. Therefore, the heating effect, and thus the temperature of the dynamic
strain aging process, must be in close conformance for both purity alloys, which implies
that the solute concentration played the major role to influence the intensity of the

dynamic strain aging.

166

 

There were situations (figures 35 and 36) when limited cross-slip of <001>
dislocations was observed for the fast cooled CP material, and dislocations were seen to
change slip planes between {110} and {001}. The activity of the mobile dislocations was
not sluggish during cross-slip as opposed to planar slip. So the intensity of DSA in
locking the diSlocations in the fast cooled commercial purity alloys was definitely not as
strong as in the slow cooled alloys. In the case of cross-slipping dislocations, the pre-
existing dislocations were of mixed types; scattered tangles of dislocations were observed
along with straight dislocations. As such, the effect of static strain aging was not also
pronounced for the fast cooled commercial purity material.

Kocks [135] proposed temporary arrest of mobile dislocations at forest
dislocations during dynamic strain aging. The solute atmosphere, which forms on the
forest dislocations, drains by pipe diffusion to the mobile dislocations during their
waiting. Thus, Kocks assumed that a dislocation may not be locked along its entire
length, but instead the solute atoms may lock only the pinned portion of the line-length at
forest dislocation junctions. This mechanism may be valid for the commercially pure
alloys in the current investigation where the presence of pre-existing tangles and their
interaction with mobile dislocations were observed. It was almost always possible to
watch and follow the movement of individual dislocation line segments. This segmental
movement may be a result of partial locking of mobile dislocation lines where the rate of
locking and unlocking is controlled by pipe diffusion of the interstitial solute atoms
during in-situ straining.

There are other theories of dynamic strain aging based upon the critical strain at

the onset of yielding and elevated temperature serrations. Those theories for high

167

 

temperature dynamic strain aging are not applicable here because the type of dynamic
strain aging witnessed in the current investigation is a low-temperature (T 5 0.17 TM)
phenomenon.

The difference in mechanical behavior can be further elucidated with
consideration of the activation area and its dependence on impurity concentration. Based
on the thermodynamic and kinetic model thermally activated dislocation motion
developed by Kocks, Argon, and Ashby [136], the apparent activation area, Aa, is related
to the Burgers vector, b by:

Aa = V*/b (14)
where Aa is defined as the area on the slip plane associated with the configuration of a

dislocation and V* is the apparent activation volume, which is related with the applied

shear stress (I), Boltzman’s constant (k), shear stain rate (1') and the temperature of

deformation (T) by:

v* = kT(6ln 7 /ar) (15)

Therefore, for a constant temperature, the activation volume depends on the

change of shear strain rate with shear stress. Kocks et al. [136] also defined the apparent
activation volume for dislocation-impurity interaction in a dilute solution as:

V* = bLAy (16)
where, L is areal density of the interstitial atoms on the slip plane, which can be
approximated as ch. Here, ‘c’ is the atomic flaction of impurities and Ay is the
activation distance over which the gliding dislocations interact with interstitial elements.

When Ay = b, equation (16) can be rewritten and the activation volume becomes:

168

 

 

v* = b3/\/c (17)

Thus, the activation volume becomes a strong function of the impurity
concentration of the deforming material. As carbon is the primary element responsible
for the strain aging behavior in MA], it is also expected to impact the activation volume
to the highest extent. Based on the total carbon concentration of the investigated alloys
(Table II), the activation volume of the commercially pure material is around 60b3 as
opposed to I76b3 for the high purity material. The values for commercially pure
materials are well in conformance with measured (59 - 60b3) and calculated activation
volumes (55b3) flom strain rate change experiments of Kitano et al. [137 and 138].
Kitano and Pollock [138] measured the activation volumes in commercially pure and
high purity NiAl single crystals. Table VIII shows the chemistry and the activation
volumes calculated for the NiAl single crystals used in the current study in comparison
with those determined by Kitano and Pollock [138]. Interestingly, Kitano and Pollock
have used the total impurity (carbon + oxygen + nitrogen + sulfur) concentration in their
calculation to determine the activation volumes, although the efficacy of any of these
elements other than carbon has been shown to be insignificant [80]. The activation
volumes are calculated to be 45b3 and 72b3 respectively for commercial and high purity
materials if total impurity levels are considered for the alloys used in the current
investigation. For such a calculation, the activation volume is assumed to be impacted in
equal proportion by carbon and other elements. However, based on the dramatic
differences observed in the dislocation behaviors of high and commercial purity materials

in the present study and in previous studies [139, 140] that is not expected to be the case.

169

 

 

Table VIII. Comparison of calculated and measured [138] activation volume for NiAl
single crystals

 

 

 

 

 

 

Material Measured Calculated V"‘b3 Calculated
Activation Vol. (Ref. 138) Activation Vol.
(V*b’) (V*b’)
(Ref. 138) (Present Study)
CP — NiAl
(210 pme and 91 59 and 60 55
ppm 0)
CP - MA]
(79 - 272 ppm C 60
and 380 ppm 0)
HP - MA]
(18 pmeand 16 83 and 90 130
- 29 ppm 9)
HP — NiAl
(32 ppm C and 160 176
ppm 0)

 

 

 

 

 

 

170

The disparity between the experimentally measured (83b3 -— 90b3) and the
theoretically calculated (130b3) values of activation volumes for the high purity material
needs some review. Two conditions may lead to this difference. There may be a
threshold limit of impurity concentration, below which the impurities are not effective in
developing efficient barriers to dislocation motions at room temperature. Two, as
proposed by Kitano and Pollock, some other intrinsic short-range barriers to dislocation
motion, such as a large Peierls stress, may exist [138]. As significant cross-slip is
observed during in-situ straining of the high purity material [139, 140], dislocation
motion may not be restricted by limited mobility of screw dislocations, as is the case for
BCC metals. The presence of a large Peierls barrier or requirement of a large shear
stress to move dislocations is not consistent with extensive cross-slip.

In spite of the difference in activation volume values, it may be concluded that the
flow behavior of the NiAl single crystals studied here was strongly influenced by
differences in the activation volumes between the high purity and the commercially pure
materials which in turn was dependent on the solute concentrations, regardless of how the
activation volume was calculated.

Difference in dislocation behavior between high purity and commercial purity
materials may also be addressed flom the lack of dislocation multiplication in the
commercial purity material by multiple cross glide mechanism. According to Gilrnan
[123], collision of mobile dislocations induces self-excited oscillations that result in
multiple cross-glide. One of the pre-condition to have self- excited oscillations is
minimum damping, which is a function of crystal purity. The damping coefficient is

higher for the commercial purity material than the high purity material and this in turn

17]

 

influences the drag forces more on the dislocations in the commercial purity material.
Also, movement of dislocations during multiple cross-glide processes is always
accompanied by collisions with randomly distributed shear-strain fluctuations on primary
and secondary planes. This kind of fluctuations has its origin in the difference between
the line tensions of the dislocations, which increases with the rate of dislocation
multiplication [123]. Therefore, with slow rate of movement in commercially pure
materials, dislocation multiplication is also slower that results in decreased plasticity than
high purity materials.

Although the bulk mechanical properties varied widely depending on the cooling
rate, the dislocation behavior during in-situ straining did not reflect the same pattern. The
in-situ straining seems to be less susceptible to changes in cooling rate or static strain
aging. The strain-aging effect observed during in-situ deformation involves both static
and dynamic strain aging. The SSA takes place during slow cooling from elevated
temperatures prior to deformation; the DSA occurs during in-situ straining. The
probability of strain aging taking place in a dynamic manner at room temperature under
constant load might be considered negligible due to the following reasons. There is not
enough thermal activation for the solute atoms to form dynamic atmospheres around the
dislocations, and it is very difficult for the solute atoms to keep up with the fleshly
generated dislocations due to limited diffusion at room temperature.

Thus, the only differentiating factor in the extent of strain aging between the two
alloys is the intensity by which the pseudo-dynamic solute atmosphere forms around the
dislocations or in their vicinity during their migration or waiting at obstacles. Therefore,

any impediment to the fleshly generated dislocations was a function of the availability of

172

 

solute atoms, the waiting time, and their diffusion. Pre-existing dislocation substructures
and the extent of their interactions during SSA do not seem to influence the mobile
dislocation characteristics, including slip processes, velocities and multiplication.

Mobile dislocation characteristics did not vary with the condition (un-deformed,
post-deformed, or in-situ) of the specimens. Dislocations generated during in-situ studies
of tensile specimens were confirmed with 3 mm disc straining. Both pre-existing and
mobile dislocations were characterized as <001> type (figure 61, 62, 65). The majority
of dislocation characterization studies in NiAl identified <001> as the mobile
dislocations, except for the cases when single crystals were oriented to the <001>
orientation. Figures 66, 68 and 69 are examples that show <101> mobile dislocations
whereas, figures 63, 64 and 67 show <lll> mobile dislocations. Mobility of <11 l>
dislocations was found in both post-deformed observation as well as in-situ strained
specimens. There are several reports of <111> dislocations in hard oriented samples [6,
7], but mobility of <1 11> dislocations was not mentioned.

Dislocations that were characterized as <11 1> screw dislocations found to move
on {112} planes in slow-cooled soft oriented high purity specimens. These obserVations
are consistent with earlier findings of Veyssiere and Noebe [47] that <111> slip in NiAl
occurs on {112} planes. In contrast with the present study however, they found
<111>{112} slip in hard oriented samples deformed at room temperature in compression.
One of the possible reasons for this difference may be the localized rotation of the tensile
axis during thin-foil straining in the present investigation. It may be recalled that the thin
area and the hole in a TEM foil create a complex state of stress. Field et al. [40] also

suggested activity of <lll> dislocation on {112} planes below the DBTT. It is

173

interesting to note that Morris et al. [102] did generate <111> dislocations in NiAl by
hardness indentation but failed to observe any mobility of the same. In hard oriented
high purity samples, <110> dislocations were found to be mobile both on {110} and
{112} planes and cross-slip was almost always evident. In order for any material to
exhibit plastic flow, it is essential that Von Mises criterion is satisfied in polycrystalline
condition. Von Mises criterion, which requires a minimum of five independent slip
systems, accommodates any arbitrary shape change. Thus, the presence of <1 1 l> mobile
dislocations has important implications on plastic deformation and mechanical behavior
of NiAl alloys.

In spite of having significant dislocation mobility and multiplication, the question
remains why NiAl ultimately fails in a brittle manner at room temperature. One of the
reasons may be the immobilization of screw dislocations. The screw dislocations were
much more mobile than the edge dislocations, and a good proportion of those screw
dislocations were eventually immobilized during in-situ straining by the development of
jogs or zigzag configuration. Brittleness results when such immobilization is not
compensated for by generation of flesh dislocations that can overcome the interactions of
the existing dislocations [141]. The brittleness of the BZ NiAl alloy systems may also be
associated with the inability to relieve the stress concentrations developed during

deformations as a result of inadequate number of operative slip systems [142].

174

 

6.0 CONCLUSIONS

In-situ straining in transmission electron microscopy has been used to characterize
the dislocation generation and motion in N iAl single crystals. Depending on the impurity
concentrations and the imposed cooling rates the behavior of the NiAl single crystals
varied. The yield stress values of the NiAl single crystals varied due to the difference in
response to the static strain aging, which was a function of the impurity concentrations in
the single crystals. The yield stress was inversely related to the final cooling rates flom
elevated temperatures. For commercially pure material the yield stress increased flom
150 MPa to 450 MPa with a decrease in cooling rate from I K/sec to 0.1 K/sec. The
influence of cooling rate was not so pronounced in the high purity material and the
maximum increase was two-fold (flom 135 MPa to 274 MPa) when the cooling was
decreased flom faster to slower rate. The differences in the yield characteristics of the
NiAl single crystals were explained by the Cottrell Atmosphere formation.

Deformed slow cooled commercially pure materials showed predominantly
tangled dislocations, whereas only a few dislocation tangles were observed in the fast
cooled. condition, reflecting less intense interaction between dislocations and solute
atoms. No discernible difference has been observed between the dislocation behaviors of
the high purity specimens depending on the final cooling rate, implying that there may be
a critical impurity concentration and/or an extreme slow cooling rate required to impart
such changes that were not provided in the current investigation.

Dislocation motion was observed in all of the NiAl samples, regardless of sample

purity and thermal treatment. However, no movement of any of the pre-existing

175

 

dislocations was observed. The rate of movement of the dislocations varied inversely
with the impurity concentration. The range of velocities of dislocations in the high purity
material was 4 —— 25 urn/sec in contrast to a range of 0.1 — 3 urn/sec for the commercially
pure material. Dislocations in the high purity material moved by quick jumps and
extensive cross-slip, in contrast to the slow and progressive movement associated with
planar slip in the commercially pure material. Cross-slip of screw dislocations was
typically followed by the formation of loops and dipole debris. The order of magnitude
higher velocity of dislocations in the high purity material indicated less severe strain
aging compared to the commercially pure material.

The difference in dislocation mobilities was found to be consistent with different
models of dynamic strain aging. The mobile dislocations could not be pinned in the high
purity single crystals because of the low concentration of solute atoms. The time
required to pin mobile dislocations in the high purity single crystals was higher than the
waiting time of the dislocations at different obstacles and thus, no significant dynamic
strain aging was observed. The waiting time was ample for the solute atoms to lock the
mobile dislocations in the commercially pure single crystals due to higher solute
concentrations. The high solute concentration in the commercially pure single crystals
resulted in enhanced damping, and dislocation multiplication could not take place at the
equivalent rate as in the high purity single crystals.

The cooling rate did not influence the dislocation substructures developed by in-
situ straining in the commercially pure material. In both slow and fast cooled samples,
dislocations were seen to move in a wiggling manner and dislocation multiplications took

place by the operation of single-ended Frank-Reed sources. In all cases, both <001> and

176

 

<110> dislocations were found to move, confirming that no significant differences were
observed in the types of mobile dislocations with cooling rate variation. As the change in
cooling rate impacts only the population of the mobile dislocations prior to in-situ
straining, the characteristics of fleshly generated dislocations did not vary significantly.

In addition to <001> and <110> mobile dislocations, mobility of <111>
dislocations was observed in both post-deformation studies, as well as during in-situ
straining of the high purity single crystals. The <111>{110} and <111>{112} slip
systems have important implications for polycrystalline deformation, since these satisfy
the Von Mises criterion requiring five or more independent slip systems to be able to
obtain a general homogeneous strain. Moreover, the flequent mobile dislocation
multiplication observed with straining increased the ability to undergo plastic
deformation in the high purity single crystals.

No particular difference in mobile dislocation behavior was observed during in-
situ straining as a result of changing the bulk tensile axes; strict adherence to the bulk
tensile axis during deformation was neither possible nor guaranteed due to the thin-foil
conditions and localized rotation of the tensile axis.

The computed local shear stress of the curved and bowed-out dislocations (50 —
67 MPa) corresponded well with the theoretical shear stress (45 — 60 MPa) of the NiAl
single crystals. Theoretically calculated (55b3) and experimentally measured (59-60b3)
values of activation volumes [118] were in close conformance for the commercially pure
material (60b3) in the present investigation. Similar validation could not be obtained for

the high purity material.

177

 

APPENDIX

_I_)_1grtal video disk of clippings flom in-situ TEM Studies

178

H

10.

ll.

12.

l3.

14.

REFERENCES

. R. Darolia, Journal of Metals, 43, 44-49, 1991.

High Temperature Ordered Intennetallic Alloys IV, MRS Proceedings, 213, editors
L. A. Johnson, D. P. Pope and J. O. Steigler, 1991.

High Temperature Aluminides and Intermetallics, Materials Science and Engineering
A, 152-153, editors S. H. Whang, D. P. Pope and C. T. Liu, 1992.

Structural Intermetallics, Proceedings of the First International Symposium, editors R.
Darolia, J. J. Lewandowski, C. T. Liu, P. L. Martin, D. B. Miracle and M. V. Nathal,
TMS, 1993.

High Temperature Ordered Intennetallic Alloys V, MRS Proceedings, 288, editors 1.
Baker, R. Darolia, J. D. Whittenberger and M. H. Yoo, 1993.

D. B. Miracle, Acta Metallurgica et Materialia, 41 (3), 649-684, 1993.

R. D. Noebe, R. R. Bowman and M. V. Nathal, International Materials Reviews, 38,
(4), 193 -232, 1993.

M. H. Yoo, S. L. Sass, C. L. Fu, M. J. Mills, D. M. Dimiduk and E. P. George, Acta
Metallurgica et Materialia, 41 (4), 987-1002, 1993.

High Temperature Ordered Interrnetallic Alloys VI, MRS Proceedings, 364, editors 1.
Baker, R. D. Noebe, D. S. Schwartz, S. Hanada, and J. A. Horton, 1995.

Deformation and Fracture of Ordered Intermetallics 111, Proceedings, editors W. O.
Soboyejo, T. S. Srivatsan and S. L. Fraser, TMS, 1996.

D. P. Pope and R. Darolia, MRS Bulletin, 21 (5), 30-36, May 1996.

High Temperature Ordered Intermetallic Alloys VII, MRS Proceedings, 460, editors
C. C. Koch, C. T. Liu, N. S. Stolof, A. Warner, 1997.

Structural Intermetallics, Proceedings of the second international symposium, editors
R. Darolia, R. Wagner, M. Yamaguchi, C. T. Liu, P. L. Martin, D. B. Miracle and M.
V. Nathal, TMS, 1997.

G. Bozzolo, C. Amador, J. Ferrante and R. D. Noebe, Scripta Metallurgica et
Materialia 33 (12), 1907-1913, 1995.

179

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

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

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

D. F. Lahrman, R. D. Field and R. Darolia, High Temperature Ordered Intermetallic
Alloys IV, MRS Proceedings, 213, editors L. A. Johnson, D. P. Pope and J. O.
Steigler 603-608, 1991.

J. E. Hack, J. M. Brzeski and R. Darolia, Scripta Metallurgica et Materialia 27 (10),
1259-1263, 1992.

J. E. Hack, J. M. Brzeski, R. Darolia and R. D. Field, High Temperature Ordered
Intermetallic Alloys IV, MRS Proceedings, 288, editors I. Baker et al., 1197-1202,
1993.

J. M. Brzeski, J. E. Hack, R. Darolia and R. D. Field, Materials Science and
Engineering A, 170, 11-18, 1993.

R. Darolia, D. Lahrman and R. Field, Scripta Metallurgica et Materialia, 26 (7), 1007-
1012, 1992. '

R. Darolia, D. Lahrman and R. Field, High Temperature Ordered Intermetallic Alloys
V, MRS Proceedings, 288, editors I. Baker, R. Darolia, J. D. Whittenberger and M.
H. Yoo, 423-428, 1993.

R. W. Margevicius and J. J. Lewandowski, Scripta Metallurgica et Materialia, 25 (9),
2017-2022, 1991.

R. W. Margevicius and J. J. Lewandowski, Acta Metallurgica et Materialia, 41 (2),
485-496, 1993.

R. W. Margevicius, I. E. Locci and J. J. Lewandowski, Structural Intermetallics,
Proceedings of the first international symposium, editors R. Darolia, J. J.
Lewandowski, C. T. Liu, P. L. Martin, D. B. Miracle and M. V. Nathal, TMS, 577-
584, 1993.

V. I. Levit, I. A. Bul, J. Hu and M. J. Kauflnan, Scripta Metallurgica et Materialia, 34
(12), 1925-1930, 1996.

T. Hughes, E. P. Lautenschlager, J. B. Cohen and J. O. Brittain, Journal of Applied
Physics, 42 (10), 3705-3716, 1971.

E-T. Henig and H. L. Lukas, Z. Metallkde, 66, 98-106, 1975.

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

180

 

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

H. Jacobi and H. J. Engell, Acta Metallurgica, l9 (3), 701-711, 1971.

R. J. Wasilewski, Transactions of the Metallurgical Society of AIME, 236, 455-457,
1966.

R. J. Wasilewski, S. R. Butler and J. E. Hanlon, Transactions of the Metallurgical
Society of AIME, 239, 1357-1364, 1967.

A. Parthasarathi and H. L. Fraser, Philosophical Magazine A, 50(1), 89-100, 1984.
P. Nagpal and 1. Baker, Metallurgical Transaction, 21A, 2281-2282, 1990.

R. J. Wasilewski, Journal of Phys. Chem. Solids, 29, 39-49, 1968.

G. W. West, Philosophical Magazine A, 9, 979-991, 1964.

X. Ren and R. Otsuka, Philosophical Magazine A, 80, 467-491, 2000.

N. Rusovic and E-T. Henig, Phys. Status Soilidi. A, 57, 529-540, 1980.

H. L. Fraser, M. H. Loretto, R. E. Smallman and R. J. Wasilewski, Philosophical
Magazine A, 32, 873-875, 1975.

R. D. Field, D. F. Lahrman, and R. Darolia, High Temperature Ordered Intermetallic

Alloys IV, MRS Proceedings, 213, editors L. A. Johnson, D. P. Pope and J. O.
Steigler, 255-260, 1991.

C. H. Lloyd and M. H. Loretto, Phys. Status Soilidi. A, 39, 163-170, 1970.

W. J. Yang, F. Lin and R. A. Dodd, Scripta Metallurgica, 12 (3), 237-241, 1978.

G. W. Marshall and J. O. Brittain, Metallurgical Transactions, 7A, 1013-1020, 1976.
D. Farkas, R. Pasianot, E. J. Savino and D. B. Miracle, High Temperature Ordered
Intermetallic Alloys IV, MRS Proceedings, 213, editors .L. A. Johnson, D. P. Pope
and J. O. Steigler, 223-228, 1991.

M. H. Loretto and R. J. Wasilewski, Philosophical Magazine A, 23, 1311-1328, 1971.
M. J. Mills and D. B. Miracle, Acta Metallurgica et Materialia, 41 (1), 85-95, 1993.

P. Veyssiere and R. Noebe, Philosophical Magazine A, 65 (1), 1-13, 1992.

W. J. Yang, R. A. Dodd and P. R. Strutt, Metallurgical Transactions, 3, 2049-2054,
1972.

181

 

49

50.
51.
52.
53.

54;

55.

56.

57.

58.

59.
60.

61.

62.
63.

64.

65.

66.

67.

M. Yamaguchi, D. P. Pope, V. Vitek and Y. Umakoshi, Philosophical Magazine A,
43(5), 1265-1275, 1981.

A. Ball, Philosophical Magazine A, 20, 113-124, 1969.
J. Bevk, R. A. Dodd and P. R. Strutt, Metallurgical Transactions, 4, 159-166, 1973.
A. Ball and R. E. Smallman, Acta Metallurgica, 14, 1517-1526, 1966.

M. H. Yoo, T. Takasugi, S. Hanada and O. Izumi, Materials Transactions JIM, 31 (6),
435-442, 1990.

R. T. Pascoe and C. W. A. Newey, Phys. Status Soilidi. A, 29, 357-366, 1968.

R. D. Field, D. F. Lahrman and R. Darolia, Acta Metallurgica et Materialia, 39 (12),
2951-2959, 1991.

H. L. Fraser, R. E. Smallman and M. H. Loretto Philosophical Magazine A, 28, 651-
665, 1973.

R. R. Bowman, R. D. Noebe, S. V. Raj and I. E. Locci, Metallurgical Transactions,
23A, 1493-1508, 1992.

E. P. Lautenschlager, T. C. Tisone and J. O. Britain, Phys. Status Soilidi. A, 20, 443-
450, 1967.

D. I. Potter, Materials Science and Engineering A, S, 201-209, 1969.

M. G. Mendiratta and C. C. Law, Journal of Materials Sciences, 22, 607-611, 1987.

J. T. Kim and R. Gibala, High Temperature Ordered Intermetallic Alloys IV, MRS
Proceedings, 213, editors L. A. Johnson, D. P. Pope and J. O. Steigler, 261-266, 1991.

1. Baker, Materials Science and Engineering A, 192/193, 1-13, 1995.

R. D. Field, D. F. Lahrman and R. Darolia, Acta Metallurgica et Materialia, 39 (12),
2961—2969, 1991.

R. T. Pascoe and C. W. A. Newey, Metal Science Journal, 2, 138-143, 1968.

K. .H. Hahn and K. Vedula, Scripta Metallurgica et Materialia, 23 (1), 7-12, 1989.
E. P. George and C. T. Liu, Journal of Materials Research, 5 (4), 754-762, 1990.

E. P. Lautenschlager, D. A. Kiewit and J. O. Britain, Transaction of AIME, 233,
1297-1302, 1965.

182

 

68.

69.

70.

71.

72.

73.

74.

75.

76.

77

78.

79.

80.

81.
82.

K. Vedula and J. R. Stephens, High Temperature Ordered Intermetallic Alloys II,
MRS Proceedings, 81, editors N. S. Stolloff, C. C. Koch, C. T. Liu and O. Izumi,
381-391, 1987.

A. G. Rozner and R. J. Wasilewski, Journal of Institute of Metals, 213, 169-175,
1966.

R. D. Noebe and R. Gibala, Scripta Metallurgica et Materialia, 20, 1635-1640, 1986.

R. D. Noebe, R. R. Bowman, C. L. Cullers and S. V. Raj, High Temperature Ordered
Intermetallic Alloys IV, MRS Proceedings, 213, editors L. A. Johnson, D. P. Pope
and J. O. Steigler, 589-596, 1991.

K-M. Chang, R. Darolia and H.‘ A. Lipsitt, High Temperature Ordered Intermetallic
Alloys IV, MRS Proceedings, 213, editors L. A. Johnson, D. P. Pope and J. O.
Steigler, 597-602, 1991.

M. H. Yoo and C. L. Fu, Scripta Metallurgica et Materialia, 25 (10), 2345-2350,
1991.

S. V. Raj, R. D. Noebe, and R. R. Bowman, Scripta Metallurgica et Materialia, 23
(12), 2049-2054, 1989.

P. Nagpal, 1. Baker, F. Liu and P. R. Munroe, High Temperature Ordered
Intermetallic Alloys IV, MRS Proceedings, 213, editors L. A. Johnson, D. P. Pope
and J. O. Steigler, 533-538, 1991.

S. Reuss and H. Vehoff, Scripta Metallurgica et Materialia, 24(6), 1021-1026, 1990.

. K. S. Kumar, S. K. Mannan and R. K. Viswanadham, Acta Metallurgica et Materialia,

40(6), 1201-1222, 1992.

K-M. Chang, R. Daroilia and H. A. Lipsitt, Acta Metallurgica et Materialia, 40 (10),
2727-2737, 1992.

K-M. Chang, R. Daroilia and H. A. Lipsitt, High Temperature Ordered Intermetallic
Alloys IV, MRS Proceedings, 213, editors L. A. Johnson, D. P. Pope and J. O.
Steigler, 597-602, 1991.

M. L. Weaver, “Investigation of Strain Aging in the Ordered Intermetallic
Compound B-NiAl”, A Dissertation, University of Florida, Gainesville, FL, 1995.

M. L. Weaver, R. D. Noebe, and M. J. Kaufman, Intermetallics, 4, 593-600, 1996.

J. M. Brzeski, J. E. Hack, and R. Darolia, High Temperature Ordered Intermetallic
Alloys VI, MRS Proceedings, 364, editors 1. Baker, R. D. Noebe, D. S. Schwartz, S.
Hanada, and J. A. Horton, 419-424, 1995.

183

83. M. L. Weaver, R. D. Noebe, J. J. Lewandowski, B. F. Oliver, and M. J. Kaufman
Materials Science and Engineering A, 192/193, 179-185, 1995.

84. D. Golberg and G. Sauthoff, Intermetallics, 4, 143-158, 1996.

85. M. L. Weaver, M. J. Kaufman and R. D. Noebe, Metallurgical Transactions, 27A,
3542-3557, 1996.

86. J. E. Hack, J. M. Brzeski, and R. Darolia, Materials Science and Engineering A,
192/193, 268-276, 1995.

87. M. L. Weaver, M. J. Kaufman and R. D. Noebe, Scripta Metallurgica et Materialia,
29,1113-1118,1993.

88. P. Wynblatt, Acta Metallurgica, 15, 1453-1460, 1967.

89. A. Lutze-Birk and H. Jacobi, Scripta Metallurgica, 9, 761-765, 1975.
90. G. F. Hancock and B. R. McDonnell, Physica Status Solidii, A, 4, 143-150, 1971.

91. R. S. Polvani, W-S. Tzeng and P. R. Strut, Metallurgical Transactions, 7 A, 33-40,
1976.

92. A. Prakash and M. J. Pool, Journal of Materials Science, 16(9), 2495-2500, 1981.
93. P. R. Strut, R. S. Polvani, and B. H. Kear, Scripta Metallurgica, 7, 949-954, 1973.
94. M. Rudy and G. Sauthoff, Materials Science and Engineering, 81, 525-530, 1986.
95. W. J. Yang and R. A. Dodd, Metal Science Journal, 7, 41-47, 1973.

96. W. R. Kane Jr., P. R. Strutt and R. A. Dudd, Transaction of AIME, 245, 1259-1267,
1969.

97. J. Doychak, J. L. Smialek and T. E. Mitchell, Metallurgical Transactions, 20A, 499
518, 1989. '

98. R. Hutchings and M. H. Loretto, Journal of Metal Sciences, 12, 503-510, 1978.
99. J. Jedlinski and S. Mrowec, Materials Science and Engineering, 87, 281-287, 1987.

100. I. Baker, P. Nagpal, S. Guha and J. A. Horton, Sixth Japanese International
Symposium on Intermetallics, The Japan Institute of Metals, Sendai, 603-607, 1991.

101. P. Nagpal, 1. Baker and J. A. Horton, Intermetallics, 2, 23-29, 1994.

184

102. M. A. Morris, J-F. Perez and R. Darolia, Philosophical Magazine A, 69 (3), 485-
506, 1994.

103. A. K. Head, P. Humble, L. M. Clarebrough, A. J. Morton and C. T. Forwood,
Defects in Crystalline Solids, 7, editors S. Amelinckx, North-Holland Publishing
Company, 1973.

104. B-C. Ng, “Examination of Dislocation Structures in B2 Stoichiometric NiAl Alloys
using TEM and SEM Techniques”, An M. S. Thesis, Michigan State University,
1997.

105. C. S. Barrett, “Structure of Metals”, McGraw-Hill, New York, 39, 1943.

106. Duke Elder, “Critical Flicker Frequency and Television or Motion Pictures”, C. V.
Mosby, 4, 602, 1968.

107. D. R. Johnson, S. M. Joslin', B. F. Oliver, R. D. Noebe and J. D. Whittenberger,
“First .Intemational Conference on Processing Materials for Properties, editors H.
Henin and T. Oki, TMS, Warrendale, PA, 865-870, 1993.

108. W. C. Leslie, editor Quench and Strain Aging, 5, MIT Press, Cambridge, MA, 1986.

109. J. D. Baird, Metallurgical Review, 16, 1-18, 1971.

110. J. Snoek, Physica, 6, 591-592, 1939.

111. J. Snoek, Physica, 8, 711-733, 1941.

112. A. H. Cottrell and B. A. Bilby, “Proceedings of the Physical Society of London”, A
62, 49-62, 1949.

113. H. E. Rosinger, G. B. Craig and W. J. Bratina, Materials Science and Engineering, 5,
163-169, 1969/70.

1 14. R. D. Noebe and A. Garg, Scripta Metallurgica et Materialia, 30, 815-820, 1994.

115. J. Friedel, “Dislocations”, Pergamon Press, Oxford, Addison-Wesley Publishing
Company Inc., Reading, MA, 1964.

116. B-C. Ng, B. Simkin and M. A. Crimp, Deformation and Fracture of Ordered
Intermetallic Materials III, TMS Proceedings, editors W. O. Soboyejo, T. S. Srivatsan
and H. L. Fraser, 337-348, 1996.

117. B-C. Ng, B. Ghosh and M. A. Crimp, Interstitial and substitutional solute effects in

interrnetallics, TMS Proceedings, editors 1. Baker, R. D. Noebe and E. P. George,
163-176, 1998.

185

118. G. E. Dieter, “Mechanical Metallurgy”, McGraw Hill, London, 1988.

119. C. Valihe, J. Douin and D. Caillard, High Temperature Ordered Intermetallic Alloys
VI, MRS Proceedings, 460, editors C. C. Koch, C. T. Liu, N. S. Stolof, and A.
Warner, 455-460, 1997.

120. Scion Image, httpz/www.scioncorp.com, Scion Corporation, Frederick, MD 217001.

121. X. Shi, T. M. Pollock, S. Mahajan and V. S. Arunachalam, High Temperature
Ordered Intermetallic Alloys VI, MRS Proceedings 460, editors C. C. Koch, C. T.
Liu, N. S. Stolof, and A. Warner, 493-498, 1997.

122. J. S. Koehler, Physical Review, 86, 52-59, 1952.

123. J. J. Gilman, Philosophical Magaazine A, 79, 329-336, 1997.

124. W. G. Johnston, Journal of Applied Physics, 33, 2716 — 2730, 1962.

125. G. T. Hahn, Acta Metallurgica, 10, 727 — 738, 1962.

126. W. G. Johnston and J. J. Gilman, Journal of Applied Physics, 30, 129-147, 1959.

127. W. G. Johnston and J. J. Gilman, Journal of Applied Physics, 31, 687 —692, 1960.

128. A. H. Cottrell, “Dislocations and Plastic Flow in Crystals”, Oxford Press at the
Clarenden, London, 1953.

129. A. H. Cottrell, “Vacancies and Other Point Defects in Metals and Alloys”, Institute
of Metals, London, 1, 1958.

130. P. G. Shewmon, “Diffusion in Solids”, TMS, Warrendale, PA, 1989.

131. J. P. Schaffer, A. Saxena, S. D. Antolovich, T. H. Sanders Jr. and S. B. Warner,
“The Science and Design of Engineering Materials”, WCB McGraw-Hill, Boston,
1999.

132. H. Oikawa, “lattice Self-Diffusion in Solid Iron; A Critical Review”, Technical
Report, Tohoku University, 46, 67-77, 1982.

133. P. G. McCormick, Acta Metallurgica, 19, 463-471, 1971.
134. P. G. McCormick, Acta Metallurgica, 20, 351-354, 1972.

135. U. F. Kocks, Metallurgical Transaction, 16A, 2109 — 2129, 1985.

186

136. U. F. Kocks, A. S. Argon and M. F. Ashby, Thermodynamics and kinetics of slip,
Progress in Materials Science, edited by B. Chalmers, J. W. Christian and T. B.
Massalski, 19, 1975.

137. K. Kitano, T. M. Pollock and R. D. Noebe, Scripta Metallurgica et Materialia, 31
(4), 397-402, 1994.

138. K. Kitano and T. M. Pollock, Structural Intermetallics, editors R. Darolia et al.,
Proceedings of the First International Conference, TMS, 591-599, 1993.

139. B. Ghosh and M. A. Crimp, High Temperature Ordered Intermetallic Alloys VII,
MRS Proceedings, 460, editors C. C. Koch, C. T. Liu, N. S. Stoloff, S. Hanada, and
A. Wanner, 467-472, 1997.

140. B. Ghosh and M. A. Crimp, Materials Science and Engineering A 239-240, 142-149,
1997.

141. X. Shi, S. Mahajan, T. M. Pollock, and V. S. Arunachalam, Philosophical Magazine
A, 79, 1555-1566, 1999.

142. V. I. Levit and M. J. Kaufman, Proceedings of the International Symposiumon

“Structural Intermetallics”, TMS, editors M. V. Nathal, R. Darolia, C. T. Liu, P. L.
Martin, D. B. Miracle, R. Wagner, and M. Yamaguchi, 683-690, 1997.

I87

    
  

 
  

' \IMilli!ljllfiiiillljjljllflllll '

3 023