Citation
A study of low and intermediate temperature slip behavior of high purity NiAl single crystals

Material Information

Title:
A study of low and intermediate temperature slip behavior of high purity NiAl single crystals
Creator:
Hu, Jian, 1962-
Publication Date:
Language:
English
Physical Description:
vii, 146 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Crystals ( jstor )
Ductility ( jstor )
High temperature ( jstor )
Room temperature ( jstor )
Single crystals ( jstor )
Specimens ( jstor )
Strain rate ( jstor )
Stress concentration ( jstor )
Temperature dependence ( jstor )
Transition temperature ( jstor )
Dissertations, Academic -- Materials Science and Engineering -- UF ( lcsh )
Materials Science and Engineering thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1997.
Bibliography:
Includes bibliographical references (leaves 140-145).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Jian Hu.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
028629806 ( ALEPH )
38855556 ( OCLC )

Downloads

This item has the following downloads:


Full Text













A STUDY OF LOW AND INTERMEDIATE TEMPERATURE SLIP
BEHAVIOR OF HIGH PURITY NiAl SINGLE CRYSTALS












By

JIAN HU


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS OF THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1997

































TO MY PARENTS

















ACKNOWLEDGMENTS


Pursuing a Ph.D degree at an age of over 30 years old and in a totally new culture and language environment presented an exciting challenge for me and this challenge could not have been overcome if I had not received the kind and invaluable guidance and advice from all of my supervisory committee members: Dr. M.J. Kaufman, Dr. F. Ebrahimi, Dr. R.T. Dehoff, Dr. R. Reed-Hill and Dr. A. Kumar. I am especially thankful to Dr. M.J.

Kaufman, my committee chairman, for his tireless technical advice and instruction, as well as for his constant encouragement to me for becoming a better person. I can never thank him enough for all his help to my growth in my career, and to my growth in other aspects of my life.

I would also like to extend my sincerest thanks to Dr. V. Levit for his countless help during the years this research project has been going on. Dr. V. Levit and his wife have been very generous in providing me with excellent experimental materials and help in conducting experiments, as well as invaluable advice and suggestions throughout this study. Without their help, the completion of this research would not have been possible. I am also very grateful to other graduate students and colleagues, especially Zheng Chen, Sanjay Shrivastava, Mark Weaver, Andy Duncan and Yongjin Lim, who constantly offered valuable suggestions and help to me.

Finally, I could never have come this far without the love and encouragement from my parents and my brothers. I am deeply indebted to them and hope this dissertation expresses my heartfelt thanks to them.















TABLE OF CONTENTS

page


ACKNOW LEDGM ENTS .................................................................................................. iii

ABSTRACT ........................................................................................................................ vi


CHAPTER

1 INTRODUCTION ..................................................................................................... I

Background ...................................................................................................................... 1

Approach ......................................................................................................................... 4

2 LITERATURE REVIEW ............................................................................................ 6

2.1 Physical Properties of NiAl ................................................................................... 6

2.2 Yield Behavior of NiAl ........................................................................................... 7

2.3 Slip Behavior ........................................................................................................ 8
2.3.1 Theoretical Calculations ................................................................................. 9
2.3.2 Observed Slip System s and Dislocation Behavior ........................................ 11
2.3.3 Effect of Heat Treatment, Alloying Elements and Impurities ........................ 12

2.4 Ductility and BDTT ............................................................................................ 13
2.4.1 Ductility of Single and Polycrystalline NiA1 ................................................. 14
2.4.2 Brittle-to-Ductile Transitions (BDT) in NiA1 ............................................... 15

2.5 Formation of Dislocation Dipoles and Loops During Plastic Deformation .......... 18

3 EXPERIM ENTAL PROCEDURE .......................................................................... 27

3.1 Sample Preparation .............................................................................................. 27
3.1.1 Single Crystal Growth ................................................................................. 27
3.1.2 Orientation Determination ............................................................................ 28
3.1.3 Heat Treatment ............................................................................................... 28
3.1.4 Preparation of Tensile Specimens ................................................................. 28
3.1.5 Preparation of Transmission Electron Microscopy Specimens ..................... 29









3.2 M echanical Testing .............................................................................................. 30

3.3 TEM Dislocation Analysis and SEM analysis ................................................... 31

4 RESULTS AND DISCUSSION .............................................................................. 33

4.1 M echanical Properties .......................................................................................... 33
4.1.1 Soft Orientation at Room Temperature ........................................................... 33
4.1.2 Soft Orientation at Non-ambient Temperatures ............................................. 36
4.1.3 Heat Treatment Effects ................................................................................. 37
4.1.4 Hard Orientation .......................................................................................... 38

4.2 Slip Behavior of Non-<001> Oriented High Purity NiAl Single Crystals at RT ..... 47
4.2.1 [557] Crystals .............................................................................................. 47
4.2.2 [123] crystals ................................................................................................. 53
4.2.3 [011] Crystals .............................................................................................. 54

4.3 The Effect of Heat Treatment on Slip Behavior at Room Temperature ................. 69

4.4 Effects of Temperature and Strain Rate on the Slip Behavior of Soft-Oriented NiA179

4.5 Slip Behavior of Hard-Oriented NiA1 Single Crystals ........................................ 93
4.5.1 Dislocations in the Specimens Tested Below 573K ...................................... 93
4.5.2 Slip Behavior at 573K ................................................................................. 94
4.5.3 Slip Behavior at Temperatures Above 573K ............................................... 98

4.6 Slip Traces and Fracture Surfaces .......................................................................... 112
4.6.1 Soft Orientation ............................................................................................... 112
4.6.2 Hard Orientation ............................................................................................. 114

4.7 General Discussion ............................................................................................... 121
4.7.1 Deformation of NiAl Single Crystals ............................................................. 121
4.7.2 Tensile Ductility of Soft Orientation ............................................................... 123
4.7.3 Tensile Ductility of Hard Orientation NiAl Crystals ....................................... 129
4.7.4 Potential of NiAI Single Crystals as High Temperature Structural Material .... 132

5 SUM M ARY AND CONCLUSIONS ........................................................................ 136

LIST OF REFERENCES ................................................................................................ 140

BIOGRAPHICAL SKETCH .......................................................................................... 146














Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

A STUDY OF LOW AND INTERMEDIATE TEMPERATURE SLIP

BEHAVIOR OF HIGH PURITY NiAl SINGLE CRYSTALS By

JIAN HU

December, 1997



Chairman: Dr. Michael J. Kaufman Major Department: Materials Science and Engineering


The slip and mechanical behavior of high purity NiAl single crystals at temperatures between 77 and 1073K has been investigated in an effort to understand their tensile properties along both non-<100> "soft" and <100> "hard" orientations and to obtain a better understanding of the observed sensitivity of the behavior to impurity elements and thermal history. It is shown that, at a strain rate of 10' s-, the high purity NiAl single crystals exhibit over 30% tensile elongation at RT when tested along certain soft orientations and the tensile elongation increases gradually from 77K (-0%) to 473K (50%) before increasing more rapidly due to the onset of dynamic recovery.

The cross slip behavior of NiAl was also studied and it is shown that, in softoriented NiAl crystals, cross slip occurs readily at room temperatures and that slip is very inhomogeneous as determined from the presence of discrete dislocation bands. A double cross slip mechanism is proposed to account for the observed high density of debris and small prismatic loops in the soft-oriented NiAl crystals after deformation and the









broadening of the dislocation bands. It is also proposed that the tensile ductility of soft NiAl crystals is controlled by the rate of cross slip and local short range diffusion which affects the development of dislocation bands and thus the local stress concentrations that ultimately lead to brittle fracture.

The slip behavior in the hard NiAl single crystals is also investigated. It is concluded that the sudden increase in tensile ductility of the crystals in this orientation is due to the activation of the <01 1>{011 } slip systems. The results of this study also indicate that, in high purity stoichiometric NiAl, <011> dislocations dominate the deformation near 573K but become unstable at and above 723K. The <011> dislocations also do not cross slip at 573K and contradictory to the predictions of theoretical calculations, these dislocations exhibit a strong tendency to lie along <11 > directions and assume a skew configuration in the slip planes. The slip behavior of hard orientation NiAl0.3Si single crystals was also studied and the results indicate that Si does not strongly affect the slip behavior of <100> oriented NiAl yet increases the tensile elongation at and above 673K by some unknown mechanisms. It is proposed that the thermally activated kink motion of <01I> dislocations along their <1I> line directions accounts for the BDTT of <001> oriented NiAI single crystals.















CHAPTER 1
INTRODUCTION


Background


In the past few decades, the potential application of intermetallic compounds in jet engines revived the worldwide interest in these materials due to the need to increase the operation temperatures of jet engines for better efficiency and the limited melting temperatures of the currently used superalloys. Many intermetallic compounds such as aluminides, silicides, carbides, and other novel intermetallics have been tested and studied in the drive to evolve usable high temperature structural materials from them. NiAl is one of the most promising and most extensively studied among these intermetallic compounds because of (1) its low density (-5.85 g/cm3), which could result in the reduction of engine weight, (2) its high melting point (1911-1955K), which is -300K higher than that of nickel, (3) its large stability range, which could ease alloying and manufacturing difficulties, (4) its good oxidation resistance, (5) its excellent electrical and thermal conductivity and (6) its low cost. However, like many other intermetallic compounds, NiAl exhibits very limited ductility and toughness at room temperature which makes it prone to handling and impact damage. Its poor high temperature strength further hinders its application as a structural material although it finds limited industry applications in areas such as coatings on turbine blades (1).

Although the yield stress of NiAI and its temperature dependence strongly depend on the orientation and various metallurgical factors, with remarkable difference between polycrystalline and single crystal forms, it normally shows characteristics resembling those









of BCC metals. That is, it has a strong temperature dependence at temperatures below about 0.15 Tm (about 300K) and a relatively temperature independent region at intermediate to high temperatures (2, 3, 4, 5, 6, 7). However, the temperature dependence of the yield stress of NiAl single crystals is reported to be strongly orientation dependent and very sensitive to both the impurity elements and deviations from stoichiometry. In single crystal form, NiAl has a much higher yield stress when oriented along <100> (cube) directions at temperatures below 600K although the high temperature yield stress is relatively insensitive to crystal orientation (3, 8, 9).

The ductility of NiAl is also strongly orientation dependent and extremely sensitive to many metallurgical parameters such as impurity concentration, and Ni/AI ratio, and is in general very sensitive to strain rate. It has been shown that polycrystalline NiA1 has virtually no ductility at room temperature (RT) (10). For single crystal NiAl, no tensile ductility has been observed in <100> oriented (hard) NiAI below -589K (4) although limited compression plasticity is obtainable (7, 8). However, in non-< 100> oriented (soft) NiAl single crystals, limited (typically <2%) tensile elongation has been obtained at temperatures at RT and below (3, 11). As temperature increases, it is reported that all forms of NiAI undergo an abrupt increase in ductility at the so-called brittle-to-ductile transition temperature (BDTT). The exact temperatures of the transitions have been found to depend on various metallurgical parameters in the case of polycrystalline NiAI as well as on the crystal orientation in the case of single crystals. This prompted many research efforts to understand the mechanisms for these transitions. However, except in the case of polycrystalline NiAI where the onset of local short circuit diffusion such as dislocation climb is generally believed to be the mechanism responsible for the abrupt increase in ductility (12), the actual mechanisms for both non-<100> and <100> orientation single crystal NiAl are still not clear.









It is suggested that some of the most important factors controlling the low temperature ductility of NiAI include the concentration of interstitial impurities in the material and the thermal history of the material. Some recent results indicated that reducing the concentration of interstitial impurities either by purifying the material or doping with gettering elements typically decrease the room temperature yield stress and sometimes result in slightly better room temperature ductility (13, 14, 15, 16). Indeed, like in the conventional BCC metals, interstitial species such as C and 0 have been shown to be responsible for the strain aging phenomena apparent in lower purity NiAl as shown in the recent work of Weaver et al. (17, 18, 19, 20) and Hack et al. (21). However, it is still not clear if the poor low temperature ductility is an intrinsic characteristic of NiA1 as there has been limited work concerning the deformation behavior in higher purity NiAl, i.e., where there is any change in the slip behavior.

The slip vectors responsible for the deformation of NiAl single crystals are <001> slip when oriented along any of the soft non-<001> orientations (7, 8, 22, 23, 24). In most reports, the observed actual slip planes are { 100} and 1 1101 depending on the orientation (i.e., Schmid factor). However, slip on higher index { hk0} planes has also been reported by Takasugi et al. (25) and by Ball and Smallman (24). The latter authors explained the apparent { hk0} slip as resulting from cross-slip between orthogonal 1 110) planes. Pascoe and Newey also reported that, at very low temperature (77K), slip is confined to { 1101 planes regardless of the crystal orientation (7) although this view is not widely accepted. Various authors (26, 27) reported that the cross-slip of <001> dislocations is relatively easy at and above room temperature. However, detailed information regarding the crossslip behavior of dislocations in NiAI such as the cross-slip plane and its effects on the ductility and other mechanical properties of NiAl single crystals is still lacking and controversial.









Because of the negligible resolved shear stress for cube slip, the slip systems in the NiAI single crystals oriented along <100> are different from those oriented along the soft directions. For the hard orientation, the results of previous studies (3, 7, 22, 27, 28, 29, 30) indicate that (1) below 600K, <111> slip dominates, (2) between 600K and 800K, there is a transition from to non- slip (e.g., <110>) and (3) above 800K, deformation is by <110> 110) slip and <001> climb. Unlike the case of the soft oriented crystals where there is no change in slip system associated with the BDTT, the sudden increase in tensile ductility of <100> crystals appears to correspond to a transition from <111> to non slip; this occurs about lOOK higher than that of the soft crystals. However, the correlation between this transition in slip behavior with the BDTT in hard NiAl crystals has not been determined, and the details of the slip behavior of <100> oriented NiAI single crystals around the temperature range where this transition occurs are still not understood, especially in stoichiometric NiAl.

It is obvious that, despite a fair amount of investigation, many details of the slip behavior in NiAl single crystals have not been studied systematically. Questions such as the role of cross-slip, the development of dislocation substructure with strain, the effect of thermal treatment on the slip behavior, the relationship between slip behavior and ductility, and the details of the orientation dependence of slip behavior still remain largely unanswered. Clearly, without a clear understanding of the intrinsic slip behavior in NiAl, any approaches to improve its ductility and toughness are difficult to justify. Therefore, a systematic study of the slip behavior in high purity NiA1 single crystals is necessary.


Approach

In this work, high purity NiAI single crystals will be tested along a few carefully chosen orientations in order to generate a reliable mechanical property database. An explanation of the mechanisms responsible for the mechanical behavior, especially for the









change in the tensile ductility of NiAl single crystals with temperature and thermalmechanical history will be attempted by performing detailed investigation of the slip and fracture behavior after testing at various conditions (temperature, strain rate, orientation, etc.). It is the purpose of this study to (1) achieve a better understanding of the slip behavior of NiAl single crystals by systematic analyses of dislocations in high purity single crystal NiAI while such factors as the impurity level, doping elements, heat treatment and specimen geometry are controlled, and (2) correlate these observations with both the deformation and fracture behavior of NiAl single crystals as a function of orientation, strain rate and temperature.















CHAPTER 2
LITERATURE REVIEW



This review will summarize previous results concerning the deformation and slip behavior of NiAl, as well as the factors affecting the deformation of NiAI in order to bring to light some important areas that need to be better understood.


2.1 Physical Properties of NiAl


NiAI is a Hume-Rothery P-phase electron compound with a B2 ordered crystal structure as shown in Figure 2.1. The stoichiometric NiAI melts congruently somewhere in the range of 1911-1955 K (31, 32). NiA1 has a relatively wide single-phase composition range (from about 45 to 60 atomic percent Ni) as indicated by the Ni-Al binary phase diagram (Figure 2.2 (33)). Unlike many other B2 compounds which disorder before melting (e.g., FeAl), NiAI remains ordered to the melting temperature consistent with its large negative heat of formation (-72 KJ/mol (34, 35)). Quenching from the vapor at rate of- 108 K/sec does not suppress the ordering of NiAl (36) so it is believed that the ordered structure is stable up to its congruent melting temperature.

The calculations and experimental measurements of the electronic band structure are in good agreement and it is generally concluded that a very strong atomic bond exists between Ni and Al along the directions, consistent with the strong ordering of this material up to the melting temperature, and much weaker bonding along the <100> directions (37). NiAl is elastic anisotropic with an anisotropy factor (A=2C44/(C1-C2)) between 3.2 (38) and 3.3 (5). This factor is only weakly temperature dependent for near-









stoichiometric compositions, and more strongly dependent on composition (38) with a trend of lower anisotropy for Al-rich and higher anisotropy for Ni-rich compositions (Figure 2.3). When the composition becomes Ni rich, an anomalous temperature dependence is observed, and this has been related to a softening of NiAl along <1 0>{ 1 10} (39, 40). Since the dislocation energies and core structures are affected by the elastic anisotropy, any change in the elastic anisotropy factor could influence the slip behavior.

It is claimed that a transfer of electrons from the Al s band to the Ni d band occurs

(41) and results in the strong Ni-Al bond energy. Consistent with such electron transfer, it is reported that off-stoichiometric binary NiAl will form constitutional vacancies (vacancies on Ni sites) when it is Al-rich and Ni anti-site point defects (Ni on Al sites) when it is Nirich, based on measurements of the composition dependence of the lattice constant and density of NiA1 (42, 43) (Figure 2.4). These constitutional defects are reported to display short-range order to avoid having like species as nearest neighbors (44) and the existence of constitutional defects makes the properties of NiAI very sensitive to any deviation from the stoichiometric composition.


2.2 Yield Behavior of NiAI


The flow stress of NiAl single crystals strongly depends on the orientation. When oriented along the "hard" <100> direction, NiAl has much higher flow stresses at low and intermediate temperatures than when oriented in all other directions. Thus the non-<100> orientations are normally referred to as the "soft" orientations while the <100> is referred to as "hard". The reported yield or flow stresses of polycrystalline NiAl alloys are much more varied probably due to the differences in grain size, processing history, interstitial content, texture and composition. The typical reported yield stress of polycrystalline NiAI has a strong temperature dependence below room temperature and above 800K, and is









relatively independent of temperature between 300 and 800K (9, 45), as shown in Figure 2.5. However, the large disagreement among the data from different investigators makes it difficult to define the details (46). The data for soft NiAl single crystals are less varied while their yield stress exhibits a very strong temperature dependence below 400K, and then remains relatively constant or slowly decreases with increasing temperature (9, 11, 24, 27) (Figure 2.6). Deviations from stoichiometry increase the yield stress of soft NiAI, while the temperature dependence appears not to be disturbed as shown by the data from a [123] oriented Ni52AI48 single crystal in Figure 2.6 (27).

The yield stress of <100> NiAI single crystals is considerably higher than that of soft crystals and display a distinctly different temperature dependence, especially below 600K (3, 9) (Figure 2.6). Specifically, there is a strong temperature dependence below room temperature and an almost constant stress region from room temperature to about 600K. Above 600K, the flow stress drops abruptly with temperature and approaches the same level as both "soft" oriented single crystals and polycrystalline NiAl above 800K. The temperature range where this abrupt decrease in yield strength takes place is strongly dependent on both composition and strain rate. For example, Kim (27) reported that, for an off-stoichiometric Ni52A148 crystal, this temperature range was about 200K higher than that of stoichiometric NiAl single crystals. A high strain rate sensitivity is reported for "hard" orientation NiAI single crystals tested at 473K and 1144K (8, 9, 11, 28). At temperatures higher than 1144K, the deformation is proposed to be dominated by the climb of <001> dislocations (47).


2.3 Slip Behavior


The slip behavior in NiA1 is an area where a lot of controversies and unanswered questions remain. This section will start with a brief summary of the theoretical efforts on the









prediction of slip systems in NiAI followed by a summary of the experimental investigations of dislocations in both "hard" and "soft" orientation NiAI single crystals.


2.3.1 Theoretical Calculations


As shown in Figure 2.1, the three shortest unit slip vectors available in the B2 structure are <001>, <110> and . These vectors represent three perfect dislocations which will retain the ordered structure while gliding. Theoretical predictions of slip systems in this class of materials requires consideration of issues such as the ordering energy, dissociation of the <111> dislocations, dislocation line energy and mobility. In B2 materials, it is well known that deformation normally occurs by either <1I> slip or <001> slip at low temperatures (48, 49). Based on the consideration that the decomposition of <111> dislocations can only occur if the separation of the <111> superpartials is small so the dislocations are compact, Rachinger and Cottrell deduced a critical value of 0.06 eV for the ordering energy (50). Below this value, the separation of < 111> superpartials is too large so that slip can only occur via the glide of <111> dislocations. This prediction agrees well with the experimental observations that slip in high ordering energy B2 compounds such as NiAl is typically via <001> dislocations, whereas in B2 compounds with lower ordering energies (e.g., FeAl), slip is normally by dislocations. However, there are some exceptions, such as AuZn and AuCd which have lower ordering energy than FeAl but still deform by <001> slip (50, 51); this may be related to other factors such as dislocation mobility.

The elastic energy per unit length of a dislocation in anisotropic cubic crystals can be expressed as (52)


= 2 * n[R
4 r r









while K is the appropriate anisotropic energy factor. Based on this equation, the dislocation line energy of all the three unit dislocations in NiAI have been calculated by Potter (51), Lloyd and Loretto (53), Ball and Smallman (2) and Miracle (54). Their results, as shown in Figure 2.7 and Table 2.1, were later verified in a more detailed calculation by Glatzel et al. (55). The relative mobility of various dislocations is also estimated using an equation proposed by Eshelby (56):
3 -2rc
S = 4 n(5exp(- -7 )
b b

The results of these calculations are shown in Table 2.1. The experimentally observed slip systems in non-<100> oriented NiAl single crystals are either <001>1110} or <001>{010}, which is in agreement with the theoretical calculations. However, in <100> NiAl crystals where there is no resolved shear stress for <001> slip, <111> dislocations are reportedly observed at temperatures below 600K and <001> and <110> dislocations above 600K. These observations contradict the theoretical predictions, which indicate that < 110> dislocations have lower line energies and higher mobilities than < 111 > dislocations.

Further efforts to estimate the dislocation mobility in NiAl have been carried out by calculating the dislocation core structures (57, 58). The core structure of <001> screw dislocations obtained with central force potentials exhibits a planar extension on 1 1101 planes (57) which predicts { 110} being the favored slip plane. However, more recent work has shown that <001>, <011> and <111> dislocations all have non-planar cores and thus should be difficult to slip (58). Parthasarathy et al. suggested that the low dislocation mobility in NiAl could be overcome by kink movement as calculations of dislocation mobility suggest that a kinked <001> dislocation in NiAI should have much higher mobility (59). Experimentally, both the dissociation and decomposition of the <110> dislocations have been observed (60) while the cores of < Ill> and < 100> dislocations in









stoichiometric NiAI are reported to be relatively compact (61, 62). Mills and Miracle (63) reported that < 110> edge dislocations exhibit a dissociated core consisting of two 1/2<11 1> superpartial dislocations, and some of the < 110> dislocations transformed into <100> dislocations during high resolution TEM observations. The dissociation or decomposition of <110> dislocations may be the cause for the strong temperature dependence of the yield stress of <001> oriented NiAI single crystals at intermediate temperatures.


Table 2.1 Calculated dislocation line energy and mobility in stoichiometric NiAl by Ball and Smallman (2) (Data in bracket by Potter (51)). * I ln( R) has been taken as being equal 4Jr
to unity.

Burgers vector Slip plane Screw/Edge E*x104 S(mobility) (ergs/cm)
[111] (-110) S 13.93(12.45) 0.982(0.735) E 29.3(26.90) 0.42(0.460)
[110] (-110) S 10.3(10.30) 0.318(0.331) E 15.25(14.78) 0.06(0.138) [010] [-101] S 9.29(9.29) 0.46(0.480) E 8.83(8.29) 0.513(0.545)
[100] [010] S 9.29(9.29) 0.67(0.653)
1 1 E 7.69(7.39) 0.71(0.718)


2.3.2 Observed Slip Systems and Dislocation Behavior


When oriented along soft directions, the observed slip systems in NiAl single crystals at room temperature and above are reported to be either <001>{1101 or <001>{010) depending on the specimen orientation, and exclusively <0014 1101 at 77K (2, 7, 8, 22, 64). Cube slip on {hk0} planes was also reported (25), however, which could be due to the composite cross slip of <001> dislocations between { 0101 and { 110) slip planes as suggested by Ball and Smallman (2). It was also reported that extensive crossslip occurs when deforming soft NiAl crystals at room temperature and above (26, 27), and









dislocation tangling, dipole and loop formation are present in the deformed specimens (24, 64).

For specimens oriented along <001>, there is no resolved shear stress on <100> dislocations so that cube slip is theoretically no longer available. At temperatures at or below room temperature, <111> slip systems were observed by Sun et al. using surface slip trace analysis and TEM (29) and by Loretto and Wasilewski (22), and Bowman et al.

(3) using TEM analysis. <001> dislocations were also frequently observed due to the kinking of compression specimens (65). <111> slip was not observed at high temperatures in stoichiometric NiAl crystals along the hard orientation and, instead, a high density of <100> dislocations were always observed. It is, however, generally believed that <110>{ 110) slip systems are active at temperatures above 600K in <001> oriented NiAl single crystals despite their infrequent observations and it has been suggested that it is the decomposition of these dislocations that produces the rather high density of cube dislocations after deformation.


2.3.3 Effect of Heat Treatment, Alloying Elements and Impurities on Mechanical Behavior


The mechanical behavior of NiA1 is well known for its sensitive dependence on impurities, alloying elements and thermal history. The effect of thermal history is related mostly to the formation of thermal vacancies introduced during the heating and cooling and therefore, the above-mentioned effects can be roughly divided into three categories: thermal vacancies, interstitial impurities, and alloying additions.

In order to improve the low temperature ductility, efforts of alloying NiAl with ternary elements have been performed in an attempt of either creating additional active slip systems (13) or gettering of interstitial impurities (16). However, no improvement in the room temperature ductility of NiAI polycrystalline alloys via alloying was reported (66) and neither was activation of additional slip systems observed via this approach (67). The only









report of enhanced ductility due to alloying was that by Darolia et al. (16) who reported that microalloying with Fe, Ga and Mo could increase the room temperature ductility of single crystal NiAl alloys, although the mechanism(s) for this improvement remain unexplained to date.

Lower purity NiAl single crystals typically exhibit higher strength and lower ductility at room temperature. It has been proposed that interstitial impurities affect the mechanical behavior of NiAl by pinning dislocations, i.e., by reducing the density of mobile dislocations. This interaction between interstitial impurities and dislocations has been confirmed recently by the observation of strain aging effects in this material (18, 19, 20).

Fast cooling from high temperature has been shown to result in supersaturation of vacancies in NiAJ (68), and, upon annealing at a lower temperature, these quench-in thermal vacancies either form dislocation loops and helices (69, 70, 71, 72), or voids (73, 74, 75, 76). While higher hardness (77) and lower ductility results from the fast cooling of NiAI from high temperatures, there has been, no investigation of the slip behavior of fast cooled NiAl single crystals and consequently, it is still not clear if the supersaturation of thermal vacancies causes any change in the slip behavior. Considering that the properties of NiA1 are also sensitive to the deviation from the stoichiometric composition, as shown by the effect of the stoichiometry on yield stress (Figure 2.8), and that such a deviation is by the creation of point defects since excess Al atoms will create vacancies on Ni sites and excess Ni atoms will create anti-site Ni on Al sites, the understanding of the effect of heat treatment on the slip behavior will also help understanding the effects of deviation from the stoichiometry.

2.4 Ductility and BDTT


The reported temperature dependence of the ductility of NiAl alloys can be roughly divided into three stages. At low temperatures, below -500K for soft orientations or









-600K for polycrystals and hard orientation single crystals, the ductility is usually quite limited as indicated by both low tensile elongation and low fracture toughness. The end of this low temperature stage is normally referred to as the brittle-to-ductile transition (BDT) temperature or BDTT, which is usually correlated with a dramatic increase in either tensile ductility or fracture toughness. At temperatures immediately above the BDTT, NiAl single crystals tend to exhibit high tensile ductilities (up to 300%) in the case of soft crystals and up to 80% in the case of hard crystals. At even higher temperatures, the elongation may drop again due to plastic instabilities (necking) in the specimen.


2.4.1 Ductility of Single and Polycrystalline NiAl


The ductility of <001> oriented NiAl is negligible except at high temperatures (>600K). For example, it has been reported that <001> oriented NiAl fails elastically in tension when tested below 589K (8, 28) although plastic deformation in compression was reported at temperatures as low as 77K (3, 27, 29). Similar to soft oriented NiAl, there is a dramatic increase in the tensile elongation of <001> crystals above certain temperatures (600K) and high elongations can be achieved at these higher temperatures (6, 28, 78) (Figure 2.9). This transition is strongly composition dependent, as shown in Figure 2.10. The actual transition temperature is near 600K with the scatter in the various data due to the variations in alloy purity, strain rate and composition (alloying additions as well as deviations from stoichiometry).

Polycrystalline NiA1 has virtually no tensile ductility (<2% typically) below about 600K. It has been suggested that the low ductility of polycrystalline NiAl is due to the insufficient number of independent slip systems provided by cube slip. Unlike single crystals, polycrystalline NiAl needs at least five independent slip systems to accommodate the grain boundary shape change according to the Von Mises criterion (79) although this requirement has been questioned by some authors (80, 81, 82).









NiAI single crystals oriented along soft orientations usually exhibit limited tensile ductility at room temperature. For example, tensile elongations of about 1 % were reported earlier for non-<001> oriented NiAI single crystals (85). However, more recent work resulted in 5-7% RT tensile elongation by adjusting the thermal treatment conditions (83, 84) or by doping with less than 0.5at% of Fe, Mo and Ga (16). Despite these improvements, the mechanism causing the low RT ductility of non-<001> oriented NiA1 single crystals is still unknown since, unlike polycrystalline materials, single crystals do-not need five independent slip systems to accommodate the shape change. Various mechanisms such as gettering or trapping of interstitial elements by solute atoms, dislocation core structure interactions, slip system modification, electronic effects and stoichiometry effects have been proposed to explain the increase in RT ductility via microalloying (16). Yet, none of them have been proven. In fact, the understanding of the mechanisms governing the low and intermediate temperature deformation of stoichiometric NiAl is so poor that the theoretical basis of such a microalloying approach is not clear and should be questioned.

As temperature increases, it is reported that soft-oriented NiA1 undergoes an abrupt brittle-to-ductile transition at intermediate temperatures and immediately above the transition temperature the tensile elongation suddenly increases to more than 100% over a short temperature range (Figure 2.9). This transition happens between 475-525K depending on the specific orientation (11, 85). The abnormally large elongation has been attributed to a balance between strain hardening and recovery processes (85). As temperature increases further, necking occurs and the plastic elongation decreases to about 20-50%.


2.4.2 Brittle-to-ductile transitions (BDT) in NiAl


Unlike the BDTT in traditional metals, the BDTT of NiAl alloys does not have a consistent definition; this makes it a rather confusing terminology and often without a clear









physical significance. In the case of single crystal NiAI, most published data about the BDTT are obtained from tensile elongation data; this may not be an appropriate representation of the BDTT. Therefore, it should be advised that the BDTT of single crystalline NiA1 may actually represent a sudden increase in tensile elongation rather than the traditionally defined onset of ductile fracture.

For stoichiometric NiAI, the BDTT of soft-oriented NiAI single crystals is reported to be between 475 and 525K depending on orientation (11, 25), ihat of <001> oriented NiA1 single crystals between 589K and 644K (28) and finally, the BDTT of polycrystalline NiAl is between 550K and 600K.

An interesting phenomenon in polycrystalline NiAI is that its BDTT (550K-600K) is higher than that of soft-oriented NiAl single crystals even though the observed active slip systems in both types of materials are the same. Since maximally only three independent slip systems are available for <001> slip, it is normally expected that polycrystalline NiAl has very limited low temperature ductility consistent with the experimental results. However, except for the few studies where non-<001> dislocations were reported after high temperature deformation, the active slip systems are still those with <001> Burgers vectors at high temperatures; this indicates that the BDTI is not due to a sudden increase in the number of independent slip systems.

Various possible mechanisms have been proposed to explain the BDT of NiAl. For polycrystalline NiAI, both the activation of new additional slip systems and climb have been suggested (53). However, the activation of new additional slip systems was not observed experimentally and neither is it consistent with the observed rate dependent deformation of NiAl near the BDTT. On the other hand, it was shown that the combination of glide and climb of <001> dislocations could result in five independent deformation mechanisms (86). Hence it has been suggested that the BDTT in polycrystalline NiAl is due to the onset of dislocation climb assisted by short circuit diffusion (66).









It was also suggested by Noebe et al. (87) that similar climb processes are responsible for the BDTT in hard-oriented NiAI single crystals since the BDTT of <001> oriented NiAI single crystals shows a similar strain rate dependence at the same temperature range as the BDTT of polycrystalline NiAI. This view was supported by the observation of a large number of <001> dislocations in hard-oriented NiAI single crystals deformed at temperatures above the BDT7. However, the BDTT of the hard-oriented NiAl is also coincident with the slip transition from <11 1> to non- slip (88). Hence, the slip system transition can not been ruled out from being responsible for the BDTT in the hardoriented crystals.

Some recent work has linked the BDTT of <001> oriented NiAl single crystals to the thermally activated mobility of <110> dislocations in that the activation of <110> slip systems was observed only at temperatures above the BDTT. Based on the observation of the decomposed core structure of <110> edge dislocations by HRTEM and the assumption that screw and mixed dislocations are relatively mobile while edge segments undergo decomposition which locks the movement of the dislocations, Mills et al. (63, 89) proposed two slightly different mechanisms all based on the diffusion-aided glide of decomposed <110> dislocations via cooperative climb of <001> dislocations as shown in Figure 2.11. However, these mechanisms rely on fast short-range diffusion which may not exist even though the local dislocation stress field and pipe diffusion along the dislocation line provide a much faster short-range diffusion than bulk diffusion at the transition temperature range. An even more severe problem with these mechanisms is that the climb of the two <001> dislocations has to be cooperative although only one of these <001> dislocations has a driving force for climb in <001> loading, which should make climb uncooperative. It is also unlikely that the decomposed <110> dislocations will constrict to their higher energy state in the presence of thermal activation. Thus, the mechanism for the BDTT of hard-









oriented NiAI single crystals still remains undetermined, although it is generally believed that some sort of thermal activation process is responsible for the transition.

The mechanism for the BDTT in soft-oriented NiAI single crystals is the one most controversial. Since the reported BDTT of soft-oriented NiAI single crystals is so low (about 0.25 TM) and the strain rate dependence is relatively small (12), the mechanisms based on the climb process mentioned above can not be readily applied. Instead, the enhanced cross-slip, or unlocking of dislocations from pinning points have been proposed to be the mechanisms responsible, in addition to dislocation climb. There are results indicating either of these mechanisms (17, 90). However, direct observations using TEM or other methods to support any of these mechanisms are still lacking. In the case of cross slip, further details of the mechanisms must also be considered since cross slip can not increase the number of independent slip systems (24). For single crystals, the Von Mises criterion no longer holds and, therefore, it has been suggested that whether the stress concentrations in the materials can be promptly relaxed determines the level of ductility of NiAI single crystals (91). This suggestion is supported by recent work which indicates that the fracture of NiAI single crystals is crack-nucleation controlled at temperatures well below the BDTT and becomes more and more crack-propagation controlled with increasing temperature (92).


2.5 Formation of Dislocation Dipoles and Loops During Plastic Deformation


It has been reported that dislocation dipoles, loops and debris are frequently observed in non-<001> oriented NiAI single crystals deformed at low to intermediate temperatures (24, 64). Similar dipoles, cusps, or loops have also been observed previously in BCC metals (e.g., in Fe-0.3Si (93)), FCC metals (94, 95), zinc (96), magnesium oxide

(97) and Si (98). These features are often attributed to mechanisms such as the pinching-off of loops from the jogs in the screw dislocations. A simple mechanism proposed by









Johnston and Gilman (99) was often used to explain the formation of dipoles and pinchedoff debris, as illustrated in Figure 2.12. This mechanism was put forward by realizing that the jogs in the screw dislocations can not glide in the same direction as the screw components. Therefore, when a jog height is small enough, the jog will be dragged to move with the dislocation leaving point defects behind. At intermediate jog heights, the sessile jog will cause dislocation dipoles to form as shown in Figure 2.12c. With superjogs, the jogged dislocation will become a single-ended dislocation source since the two edge segments can pass each other without being trapped. High densities of prismatic loops and dipoles can be readily produced via the Johnston and Gilman mechanism if the jogs can be formed via double cross slip. Arguing that the Johnston and Gilman model is too simple to account for the various and sometimes very complicated dislocation features observed in BCC and FCC metals, and noting that in many cases cross slip is unlikely unless local stress concentrations are present, Tetelman proposed another mechanism for the formation of dislocation dipoles and loops (100) as shown in Figure 2.13. In this mechanism, two dislocations of equal Burgers vector but opposite sign, slipping on parallel slip planes separated by a distance y, can trap each other by their elastic energy fields. The dislocations can then lower their energy by reorienting part of their length, and the screw segments can cross slip to annihilate and form a dipole or closed loop. This mechanism may account for the jogs and dipoles forming at the early stages of deformation where slip is confined to the primary slip planes. The jogs can also be formed via dislocation-vacancy and dislocation-dislocation interactions. However, the jogs produced by these interactions will be of very small height, and may only produce point defects when they are moving. Nevertheless, most jog forming mechanisms include a certain cross slip process. Therefore, an investigation of the details of both dipole and prismatic loop formation may result in a further understanding of the slip behavior such as the primary and cross slip planes or the intensity of cross slip during deformation.














I


[011] [100]l


Figure 2.1 The crystal structure of NiAl with the three possible slip directions shown.


WEIGHT PERCENT NICKEL


1800 . . . . . ..C',,S ..
1600 L
14550C
1400 AMN

O(NI)
1200

~1000



600
:R-'(AI)-AN1
400 -1
0 10 20 30 40 50 60 70 80 90 100 Al ATOMIC PERCENT NICKEL Ni

Figure 2.2 Binary NiAI phase diagram (from Nash, Singleton, and Murray (33)).










250




200



") 150
0
o

0
rnl00


45 50 55 60 Ni (at%)

Figure 2.3 The elastic constants of NiAI as a function of temperature and composition. (after Rusovic and Warlimont (38)).


0-289 0288 0287 0-286


density vw
""


50 .60


7


6 .) mD


Ni (%)

Figure 2.4 Lattice parameter and density of NiAI as a function of composition at room temperature (after (12)).











2000 1800

1600 o 1400 0
"- 1200 1000



.9 600


0 200 400 600 800 1000 1200 1400 Temperature (K)


Figure 2.5 Temperature dependence of yield stress of polycrystalline NiAl (after (12)).


2000 1800 1600

1400 1200 1000 800 600


400 "Lj Y /

200[ [ii-T
0 Y

0 200 400 600 800 1000 1200 1400

Temperature (K) Figure 2.6 Temperature dependence of yield stress of NiAl single crystals for different orientations (Data compilation from (12)).













3.5 1*".

3 [ 100](011) c[011I(01 1)
2.5o [111J(011)





- 1.5




0.5


0 135180

Theta (degrees)


Figure 2.7 Line energy of [100], [011], and [11] dislocations as a function of dislocation character on the (011) plane at 933K. The horizontal axis is the angle between the dislocation line vector and the [ 100] dislocation. The line energy is proportional to C44(a2/4n)ln(r/ro).(after Miracle (54)).


1000 900 800

0- 700

600
500

C)
400

._ 300

200 100

0
46 50 54 58 62 Ni (at%)

Figure 2.8 Yield stress of NiAl as function of composition (Data compilation from (12)).


















1202
= 122)
0
0
80 -001)


400




01
0 200 400 600 800 1000 1200
Temperature (K) Figure 2.9 Tensile elongation vs. temperature along different orientations ((6)).


200 400 600 800


1000 1200


Temperature (K) Figure 2.10 Tensile ductility of <001> oriented Ni-50AI and Ni-60AI single crystals as a function of temperature (after (12)).





















40O 1 0[ Extra (0 2 2) half-plane ,, y


z x
[1 0] [of[ 1


Vacancy


Motion
"Diffusion Path" Motio Extra (0 2 2)
0 half-plane

400 i]


I1


(011) Glide
Plane C(


Figure 2.11 Mechanisms by Mills et al. (63, 89, 101) for the glide of decomposed <110> dislocations: (a) Cooperative motion of a decomposed [011] dislocation via a short-range diffusive process between two <001> dislocations. Diffusion of atoms from the tip of the lower extra (022) plane to the upper one accomplishes the glide motion of the overall [011] dislocations. (b) Forward propagation of a decomposed edge segment by the lateral motion of macro-kinks (MK). Glide motion of the MK requires constriction of the decomposed segments, which can occur by a conservative climb process.


Overall

























(C)
Figure 2.12 A model explaining the behavior of jogged screw dislocations by Gilman and Johnston: (a) small jog is dragged along creating point defects as it moves; (b) in the case of very large jog, segment NY and MX move independently; (c) intermediate jogs (MN) pinning dislocations and the segments NP and MO interact and can not pass each other in a process forming edge dipoles. (Figure from Low and Turkalo (93))


Figure 2.13 The mechanism for jog formation proposed by Tetelman: (a) two non-parallel dislocations MM' and NN' of opposite sign glide on parallel slip planes separated by a distance of y; (b) dislocations lower their energy by reorienting part of their lengths in the glide plane; (c) PM' cross slips down and leave a dislocation dipole MPP'R'RN' after annihilating the cross slip segment. (after Hull (102))















CHAPTER 3
EXPERIMENTAL PROCEDURE


3.1 Sample Preparation



3.1.1 Single Crystal Growth


Stoichiometric (Ni50Al50) NiAI and a 0.3at%Si-doped NiAI were chosen for the purpose of this study. High purity Ni (99.99%), Al (99.999%) and Si (99.9999%) were used as starting materials. The materials were first cast into rods using non-consumable arc melting under purified argon. These arc-melted rods were then used to produce single crystal rods about 27 mm in diameter and 80 mm in length in a CENTORR Model M60 furnace under purified Ar atmosphere using a modified Bridgman growth method. High purity Al203 crucibles were used in order to minimize contamination. After growth, the single crystals were homogenized at 1573K for 3 hours and slowly cooled to room temperature. All the specimens studied in this research are as-homogenized unless indicated otherwise.


Table 3.1 Chemical composition of single crystals used in this study.
Crystal Orientation Ni Al Si C S 0 N
ID at. % at. % ppm ppm ppm ppm ppm
62 [001] 49.6 50.3 460 19 -
34 [557],[233] 50.4 49.5 - 82 <5 106 <15
[011]
109 [557] 50.3 49.6 - 24 - 120 <15 64 [011] 50.6 49.3 240 180 - 81 <15
40 [123] 50.0 49.9 100 110 13 100 9
NiAI-Si [0011 50.28 49.44 0.25 at. % 151 <7 89 15









The chemical composition of the final single crystal rods were analyzed at NASA Lewis Research Center by analytical wet chemistry/titration techniques (Ni and Al), ultraviolet/visible spectrophotometry (Si) and combustion techniques (0, C, S, N). The results of these chemical analyses are shown in Table 3.1.


3.1.2 Orientation Determination


The orientation of the as-grown single crystal rods and some of the tensile specimens before and after deformation were determined using the back reflection Laue technique. Precautions were taken to ensure high accuracy of the specimen orientation. In most cases, the deviation of the actual specimen was less than 2 degrees from the intended orientation.


3.1.3 Heat Treatment


All the heat treatments were carried out in tube furnaces under purified flowing argon. The samples were wrapped in Ta or Mo foils before being placed in the furnace. Water quenching of the specimens was performed by quickly dragging the specimens out of the hot zone and dropping them into water. Furnace cooling was performed by leaving the specimen inside the furnace after the heat treatment and turning the power off. It is estimated that the average cooling rate was -0. 1 K/s, with slightly higher cooling rates at the higher temperatures.


3.1.4 Preparation of Tensile Specimens


Tensile specimens were cut along selected orientations using an electrical discharge machine (EDM). The original geometry and size (gage cross section about 1.6mm x 2.4mm) of the specimens are shown in Figure 3. 1 a (unless otherwise described). After EDM cutting, the specimens were electropolished in a 10% perchloric acid-90%methanol









solution and the final gage cross section was approximately 1.3mm x 2.1mm after electropolishing to remove the recast layer and any surface roughness.

The orientations (i.e., tensile axes) of specimens investigated in this study include [557], [223], [123], [011] and [001], as shown in Figure 3.lb. In most cases, the Side (narrow) and Face (wide) (Figure 3. la) were also carefully chosen to be parallel to certain crystal planes (see Table 3.2).


Table 3.2 The crystal orientations of the specimens. F(1 10) and S(l 10) refer to the cases when the Face and Side planes are parallel to (110), respectively. Specimens Orientation Side Plane Face Plane [557] NiAI S(l 10) [5571 (110) (7710) [557] NiAl F(110) [557] (77 10) (110) [123] NiAl [123] n/a n/a [233] NiAl [233] n/a n/a [011] NiAl F(110) [011] (100) (011) [011] NiA1 S(110) [011] (011) (100) [001] NiAI [001] (010) (100) [001] NiAI-Si [001] (110) (110)


3.1.5 Preparation of Transmission Electron Microscopy Specimens


Foils for TEM analysis were cut with a low speed SiC abrasive wheel from the specimen gage sections either parallel to the Side or to the Face or parallel to a certain crystallographic plane in the specimen. In order to best preserve the dislocations in the specimens, TEM foils were carefully hand ground on fine grinding papers to a finish










thickness of - 150 pm. Final thinning of the foils was done by jet-polishing using a Struers Tenupol-2 device in a solution of 70% ethanol, 14% distilled water, 10% butylcellusolve and 6% perchloric acid, at 0�C and 25V.

Other precautions such as cutting TEM foils from different regions in the gage of the tensile specimens and using a low jet speed during electropolishing were employed in order to avoid specimen damage.


3.2 Mechanical Testing


Tensile testing was carried out on a screw-driven Instron Model 1125 machine with a holder designed to accept the geometry of the specimens in this study. The specimens were heated using a clamshell furnace with an accuracy �5K when tested between RT and 1273K in air. In the case of testing at 77K or 180K, the specimen and the specimen holder were either immersed in liquid nitrogen or liquid nitrogen-alcohol mixtures, and test temperatures between 77-180K were achieved by keeping specimens in the cold N2 vapor. The temperatures were measured using a chromel-alumel thermocouple attached directly to the specimen. When testing at temperatures other than room temperature, a certain holding time was given after reaching the test temperature to ensure thermal equilibrium during testing. After deformation at elevated temperatures, the furnace was always promptly removed to ensure fast cooling of the specimens and retention of the dislocation substructures.

The shear stress (r) and shear strain (y) were calculated using the following equations (103):



P 2 1 . 2
T= cosXol0 0" sin













I- -sin A0 -CosA
cosxot L 0 o 0



where P is the load, A0 the initial cross sectional area, L, and L0 the current and initial gage lengths respectively, ?. the initial angle between the tensile axis (TA) and Burgers vector, and X0 the angle between TA and slip plane.

Due to the semi-brittle nature of NiAl crystals, normally at least two tests were performed for each data point at elevated temperatures. While at lower temperatures where larger scattering is expected, more than two tests were often conducted.


3.3 TEM Dislocation Analysis and SEM analysis


The dislocations in the deformed specimens were investigated using a JEOL200CX TEM and a Philips EM420 TEM both equipped with double-tilt stages allowing �60 (main) and �30 (cup) tilts and operating at accelerating voltages of 200 kV and 120k, respectively. Conventional bright field (BF) and weak beam dark field (WBDF) images were used to determine the nature (Burgers vector and line direction) of the various dislocations. It was found that the dislocations obey the gob criterion despite the elastic anisotropy of NiAl (53). However, whenever there was any doubt about the residual contrast at gob=0, additional diffraction conditions were used to verify the analysis.

The dislocation line directions were identified via trace analysis using standard techniques. Similarly, images from at least three zones were used to verify the results.

A JEOL-35CF SEM was used to analyze the fracture surfaces and slip traces on the gage surfaces. Again, the latter were correlated with the dislocation analyses by performing surface trace analysis.





























25 S~ide Faceoi-100


Figure 3.1 (a) The geometry and the size (in mm) of a typical tensile specimen after EDM cutting and before electropolishing. (b) Stereographic projection with the orientations investigated in this study shown.















CHAPTER 4
RESULTS AND DISCUSSION



In this chapter the results of the mechanical properties of both soft and hard orientation high purity NiAI single crystals are presented. In order to understand the mechanisms of the mechanical behavior, the slip behavior was studied using TEM and SEM. The effects of thermal treatment, temperature, strain rate and orientation on the deformation is also presented. The last section consists of a general discussion of the cross slip behavior, dislocation mobility, tensile ductility and other deformation behavior of NiAI single crystals based on the understanding obtained through this work and previous research.


4.1 Mechanical Properties



4.1.1 Soft Orientation at Room Temperature


The critical resolved shear stress (CRSS), work hardening rate, and tensile elongation of a series of "soft" single crystal specimens tested at room temperature are listed in Table 4.1. The results indicate: (1) that up to 34% tensile elongation can be achieved at RT; (2) that tensile elongation is dependent on orientation and specimen geometry; (3) that there are no significant differences in the CRSS between the <001> (110) and <001> 1010} slip systems for samples cut from the same crystal (the differences between the average CRSS value for each system is no more than 5% which is within the scatter); (4) the work hardening rate (WHR) is very low for orientations other










than <110>; and (5) the fracture stresses ((T) of all the specimens are close to each other except for that of the [ 123] oriented specimens which were cut from a different crystal. Furthermore, the tensile elongation also appears to be dependent on geometry, i.e., [557] specimens with the F( 110) geometry (see Chapter 3) tend to exhibit higher tensile elongations than those with the S(110) geometry. This difference in tensile elongation could be caused by the difference in the specimen head constraints due to the asymmetric rectangular specimen shape (Figure 4.1), as described elsewhere (104). However, it is also possible that this difference is related to the higher CRSS of the S(l 10) specimens as indicated in the table. It should be noted that although the specimens of both geometries were cut from the same crystal, the F( 110) specimens were actually cut from the bottom half of the crystal rod whereas the S( 110) samples were cut from the top half. It was subsequently determined via chemical analysis that the compositions of these crystals were not uniform throughout the crystal, i.e., the top is normally Ni-rich due to the evaporation of Al during single crystal growth; this resulted in CRSS values about 15% higher in the [557]S(1 10) specimens. Since in lower purity NiAl crystals the yield stress is normally higher and elongation normally lower, the higher CRSS in S(l10) specimens should adversely affect the tensile elongation.

Typical shear stress-shear strain relationships are shown in Figure 4.2a. Once again, it can be seen that the WHR of [557]-oriented single crystals is quite low throughout the deformation. The WHR of [123]-oriented specimens is very close to that of [557] oriented specimens suggesting similar slip behavior for both [557] and [123] oriented specimens, while that of the [011] oriented crystals is considerably higher, consistent with their double slip behavior. The [123] crystal also exhibits a lower CRSS for [001](]10) slip than that of [557] specimens probably due to its more stoichiometric composition (Table 3. 1). The RT shear stress and shear strain relationships of these specimens suggest









that the Stage I deformation for [557] and [123] oriented crystals lasts until fracture, while [011] oriented specimens undergo stage II deformation from the onset of plastic deformation. The effect of purity on the mechanical behavior is clearly shown (Figure 4.2b) by a comparison between the results from the [557] crystals in this study with those of a previous study (64) from lower purity crystals along a similar orientation ([111]). It is obvious that the higher purity crystals used in this study exhibit higher RT ductility, lower yield stress and lower WHR.


Table 4.1 Tensile data for soft-oriented NiAl single crystals along different orientations. S=S(110), F=F(I 10), B denotes to the bottom half and T the top half of a crystal rod, and
0 is the Schmid factor for the most favored slip system(s). For the [011] orientation, the CRSS (T02) is calculated for <001>{010} slip while in all other cases [001](110) slip is assumed. A 10' s-1 strain rate was used in all the tests. Crystal Orient. U 2 2 WHR 9f El. ID MPa MPa (O=d'u/dy) MPa % MPa
34B [557]F .5 114 57 62 266 34 34B [557]F .5 115 57.5 218 18 34B [557]F .5 109 54.5 190 16 34B [233] .485 117 56.5 186 5 34B [233] .485 119 57.5 202 4 34B [233] .485 127 61.5 169 3 34T [557]S .5 137 68.5 102 239 15 34T [557]S .5 134 67 221 13 34T [557]S .5 127 63.5 196 10 34T [011] .5 147 73.5 205 3 34T [011] .5 133 66.5 180 3 34T [011] .5 139 69.5 192 2 64 [011]F .5 111 55.5 468 242 7 64 [011IF .5 117 58.5 563 237 6 64 [011JF .5 102 51 230 5 64 [01I]S .5 111 55.5 448 235 6 64 [011IS .5 102 51 476 212 5 40 [123] .454 116 52.5 110 306 19 40 [123] .454 110 50 114 304 15 40 [123] .454 124 56.5 109 318 14









4.1.1 Soft Orientation at Non-ambient Temperatures


Figure 4.3 shows the temperature dependence of the CRSS ('t02) of different NiAl single crystals oriented along various soft orientations. It can be seen that the CRSS of [557] oriented NiAI single crystals exhibits a strong temperature dependence below RT and decreases only gradually with increasing temperature above RT. The low temperature (below RT) CRSS of NiAI single crystals along other soft orientations was not measured. However, it is believed that the same trend is true for all the soft-oriented NiAI single crystals based on the similarities at higher temperatures. This kind of strong temperature dependence normally indicates that thermal activation is required in the movement of dislocations, such as in the case of BCC metals. It has been shown in several studies that the Arrhenius representation T02 = to exp(-Q/RT) can be successfully used to evaluate possible thermal activation processes in NiAI (46, 87). Therefore, the results in Figure 4.3 are re-plotted in an Arrhenius plot in Figure 4.4. As is apparent in this plot, for a strain rate of 10-4 Sl, the deformation below 500K occurs by a similar process whereas, above 500K, the deformation process becomes more complicated.

As shown in Table 4.1, [557]F(1 10) specimens have a RT tensile elongation between 16-34% when tested using a strain rate of -10. s'. However, this elongation reduced to 1.3-6% when tested at 120K and increased to more than 100% when tested above 573K (Figure 4.5 and Figure 4.6). Similar to the data reported by Lahrman et al.

(11) and Takasugi et al. (6), elongations of [557]F(1 10) specimens reach a peak value at intermediate temperatures (Figure 4.5). It is also noted that the magnitude and temperature of the peak elongation depend on the strain rate. As shown in Figure 4.5a, the peak temperature was shifted to higher temperatures with increasing strain rate.

At 473K and 1073K, the effect of strain rate on elongation actually reverses (Figure 4.5b). At 473K where plastic instability (i.e., necking) does not occur, or 573K where










plastic instability has just occurred, the elongation decreases with increasing strain rate; this is also true for temperatures below 473K, while at higher temperatures such as 1073K where the specimens neck easily, higher strain rates result in higher tensile elongations. However, the situation where plastic instability occurs should be considered independently from that where no plastic instability takes place. Therefore, it can be concluded that at temperatures at or below 473K, the tensile elongation of soft-oriented NiAl single crystals gradually increases with increasing temperature or decreasing strain rate, i.e., there is no abrupt change in tensile ductility. At temperatures at or above 573K, high elongations can be obtained and this high elongation is sensitive to the strain rate, consistent with the assumption that the balance between recovery processes and work hardening results in these abnormally high elongations.


4.1.3 Heat Treatment Effects


It has been shown that the mechanical properties of NiAl are sensitive to the thermal or processing history. In this work the effect of thermal treatment is studied using high purity NiAl single crystals oriented along [557]. The single slip orientation was chosen to exclude the complicated situation caused by the interactions between different slip systems. The specimens were water quenched (WQ) or furnace cooled (FC) after having undergone the same homogenization treatment followed by reheating to 1273K. To exclude the possible effects due to specimen geometry, all the specimens tested are either with the F(! 10) or S( 110) geometries (see Chapter 3) and only the mechanical behavior of the specimens of the same geometry were used for comparison.

The RT mechanical properties of [557] oriented NiAI single crystals are listed in Table 4.2 where it is apparent that the cooling rate greatly affected the RT yield stress and tensile elongation. The CRSS of the WQ specimens was almost double that of the FC specimens, whereas the elongation decreased from about 10% in the case of FC specimens










to less than 1.5% in the case of WQ specimens. It is believed that the substantial decrease

in the tensile elongation and increase in the CRSS of WQ specimens are due to their higher

thermal vacancy concentrations.




Table 4.2 Influence of thermal treatment on the tensile properties of NiAI. Specimens are [557]F(1 10) single crystals. FC=furnace cooled; WQ=water quenched. Note that the WHR for WQ specimens is estimated with data at small strains where the deformation is still in the transition range before reaching Stage I deformation. Therefore, it may not represent the Stage I value. * based on the data in Table 4.4.1.
Heat 'CO 2 5f (MPa) WHR Elongation
Treatment MPa (0=dt/dy) % MPa
1273K, FC 58.5 185 -100* 10 1273K, FC 56 173 7 1273K, FC 60 191 10
1273K, WQ 112.5 244 -300 1.5
1273K, WQ 121.5 273 1
1273K, WQ 178 <0.1


4.1.4 Hard Orientation


Figure 4.7 shows the temperature dependence of the yield stress for high purity NiA1 tested along the <001> "hard" orientation in tension. The yield stress is above 600 MPa at temperatures below 573K and decreases rapidly with increasing temperature above 573K. Doping with 0.3 at% Si appears to increase the yield stress at temperatures below 773K and to produce a more abrupt drop between 573 and 773K. However, at temperatures above 773K, both NiAl and NiAI-Si specimens have similar low yield stresses.

Below 473K, the tensile elongations of the specimens are typically less than 0.5% and, in many cases, the specimens fractured before yielding. The specimens typically have

-2% tensile elongation at 473K and 523K, and undergo more than 20% elongation before









fracture at and above 573K. Therefore, both the NiA1 and NiAI-0.3Si single crystals tested in this work appear to undergo a rather sudden increase in ductility above 473K. The tensile elongation becomes relatively temperature independent above 773K for binary NiAl whereas the Si-doped crystals exhibit considerably higher elongations at temperatures above 673K.

In order to reveal possible transitions in the thermal activation of the slip behavior, the yield stress shown in Figure 4.7 is re-plotted in an Arrhenius representation in Figure 4.9. It is obvious that a transition occurs, similar to that reported previously (46). However, the transition in the high purity NiAI appears to occur at lower temperatures and be more gradual in nature. In both Figure 4.7 and Figure 4.9, the Si-doped crystals appear to exhibit a higher transition temperature than that of the binary crystals even though this difference was not reflected by the temperature dependence of the tensile elongation in Figure 4.8.

From the results of this study, it appears that the sudden increase in ductility coincides with a drop in yield stress, and that this drop in yield stress is more gradual than what was previously reported (3, 9). These characteristics indicate that a thermal activation process, rather than a rigid slip system transition, may be responsible for the yield drop and ductility increase in <001> oriented NiA1 single crystals tested at intermediate temperatures. Assuming hard-oriented NiAI single crystals deform by <011> {011 } slip, the temperature dependence of the CRSS in both hard and soft orientations can also be plotted as in Figure 4.10; this shows that, similar to the results of previous researchers (12), the CRSS of hardoriented crystals drops to values approaching those of soft-oriented NiAI single crystals at high temperatures.





































Constrained (C) Less Constrained (LC)

Figure 4.1 Schematics of the rectangular tensile specimens before and after deformation. Note the shape change (shaded) is consistent with "plane strain" in that there is no change in the thickness (T) in the constrained (C) condition which corresponds to [557]S( 110) geometry, or the width (W) in the less constrained (LC) specimens which corresponds to [557]F( 110) geometry. TA is the tensile axis, b the Burgers vector and n the slip plane normal.


































0.2 0.4 0.6
Shear Strain

(a)


400 350


300

v 250 200

150 '- 100


0.05


0.15


True strain


(b)
Figure 4.2 (a) RT shear stress vs. shear strain for NiAl single crystals having different orientations obtained using a strain rate of 104 s'. (b) True stress vs. true strain for [557] crystals in this study along with the results from [111] crystals by Field et al. (64).


-0-[5571


0


6 Field-91 [111]











400 350

300 - [557]
--- -- [0 111
250 - [1231

200
U,

o 150

100

50

0
73 273 473 673 873 1073 1273
Temperature (K) Figure 4.3 Temperature dependence of the CRSS at a strain rate of 10-4 s for [557] and [123] oriented crystals ([001](110) slip system) and for [011] oriented crystals ([001 ](010) slip system).

1000





100
0.


- [5571 10 ,- -- 011]
-5--f7 11





1
0 20 40 60 80 100 120 140
1 000/Tem peratu re(K) Figure 4.4 Arrhenius representation of the temperature dependence of the CRSS for softoriented crystals.







43




200
--- SR 10-2
180 M-SR 10-3

0 160 - SR 10-4

0 140 . 120 2 100 _ 80

- 60 - 40

20

0
373 573 773 973 1173

Temperature (K)

(a)


180

160 - 473K
e- --M- 573K
140 A 1073K
.o 120 100
C 0


= 60

c 40
I
20

0
1.OOE-02 1.OOE-03 1.OOE-04 Strain rate (1/s)


(b)
Figure 4.5 Tensile elongation vs. temperature (a) and strain rate (b) for [557] specimens tested over a range of intermediate to high temperatures. SR=strain rate.












80 70

" 60

C
0
, 50 5 40
w
* 30 "20 I

10

0


100 200 300 400
Temperature (K)


Figure 4.6 Tensile elongation of [557]F( 110) NiAI single crystals at low to intermediate temperatures tested at a strain rate of 10 4S-1.


1200 1000

800



.? 600


400


200


0


373


573


773


973


1173


Temperature (K)


Figure 4.7 Temperature dependence of tensile yield stress of <100> oriented NiAl single crystals. NiA1-Si was doped with 0.3 at% Si. It should be noted that at temperatures below 573K, the data may represent the fracture stress rather than the yield stress because the activation of or <110> slip systems could not be verified by dislocation analysis.


* []

*NiAI
* U] NiAI-Si [100]




* U]
U �[ !U












80 70 60
C
0

CD o 40
w
30
0 20
0

0

273


473


673


873


1073


Temperature (K)

Figure 4.8 Temperature dependence of tensile elongation of <100> oriented NiAl single crystals.


10000




CL
1000



W

-100


5 10 15 20 25 30 35 40 10000/Temperature(K)


Figure 4.9 Arrhenius representation of the temperature dependence of the yield stress of <100> oriented NiAl single crystals.


*NiAI [1001 [ ENiAI-Si [100] U



U]

I


- NiAI [] NiAl-Si[10

























600


500

*<011>{011}
-. 400 0 <001>{1 10}
0

300
U)
n

200


100 U



0 200 400 600 800 1000 1200 1400 Temperature (K) Figure 4.10 Comparison of the temperature dependence of the CRSS of [001] and [557] oriented NiAI single crystals. The <0 11>{0111 slip systems are assumed in calculating the CRSS for the [001] orientation.









4.2 Slip Behavior of Non-<00 1> Oriented High Purity NiAl Single Crystals at RT



4.2.1 [5571 Crystals


The [557] orientation was selected for the initial investigation of the slip behavior in high purity NiAI single crystals due to the expected single slip ([001](110)) nature corresponding to this orientation. At first, the microstructure of the as-homogenized crystals was checked and as shown in Figure 4.11, the initial density of dislocations in the undeformed crystals is extremely low. In order to determine the slip system and fundamental slip behavior, a [557]S(1 10) specimen was deformed to -3% strain and foils for TEM analysis were cut both parallel and perpendicular to the active (110) slip plane. Figure 4.12 shows the dislocation substructure in the foils cut parallel to (110) along with the dislocations. It can be seen that the substructure is not uniform with tangled dislocations in some regions, which, nevertheless, indicates the easy glide of dislocations in this specimen. It will be shown later in this section that this kind of tangled substructure is actually parts of three dimensional patch-like discrete dislocation bands as illustrated in Figure 4.19. It should be pointed out that the dislocation bands in this study refer to the above-mentioned patch-like dislocation substructure. The results of g.b analysis indicate that essentially all of the dislocations have the expected [0011 Burgers vector and that little if any secondary slip occurred. Although most dislocations have a mixed character, the trace analysis shows that the average line direction is approximately [ 110] especially in the tangled regions. This is the edge direction for [001] dislocations on ( 10) planes.

Two primary features of the dislocation substructures were characteristic in these specimens. First, most dislocations are bowed out and tend to have a mixed character about 40-50' from the edge orientation, while almost no screw or close-to-screw segments were observed except for some very short segments (Figure 4.12). Second, in some









regions, the dislocations are heavily pinned and form edge dipoles sometimes leaving traces of pinched-off loops often elongated along the [110] direction (arrow points, Figure 4.12). These dislocation features are important in understanding the dislocation behavior in NiA1 as will be discussed below.

Figure 4.12a shows that the bowed-out dislocations tend to have sharply bent tips. The sharp bent feature of dislocations in NiAI was attributed by Loretto and Wasilewski

(22) to the tendency of dislocations to avoid pure screw orientations due to the relatively higher elastic energies in screw orientations. However, it is unlikely that a long screw segment could relax into much longer 40-50' segments during unloading since there is very little if any decrease in the total energy by going through such relaxation. This can be seen by calculating the total energy change AE of a unit length of screw dislocation relaxing into segments with an angle of 0 between the line direction and the Burgers vector (Figure

4.13a)
AE = E(O) * L(O) - E(O) 9 L(O) = E(6) -E(0) cos 0
while E(O) is the energy of a screw dislocation, E(O) the energy per unit length of the relaxed segments, which are shown in Figure 4.13. A preference of [001] dislocations to lie along directions (55' or 1250 from [0011) was also reported by Kim (27) and was attributed to dislocation line relaxation because these directions lie close to the theoretically-calculated, lowest energy configuration (2, 105) as shown in Figure 4.13. Nevertheless, the bowed-out dislocations with the sharp tip in Figure 4.12 could also be due to either the differences in the mobility of different [00 11 segments or pinning by small jogs at the screw segments. As illustrated in Figure 4.14, the dislocations bow out under stress and if the screw segment has the highest mobility, a tip and 40-50* segments could form. It has been suggested by various workers that the <001> screw dislocations have higher mobility than <001> edge dislocations in NiA1, especially at low temperatures,









based on the fact that <001> screw dislocations have never been observed. In fact, theoretical calculations of dislocation mobility using Eshelby's equation (56) and core structure using embedded atom methods (EAM) all suggest that <001> screw dislocations have higher mobility (106).

The other features such as the zigzag nature of some dislocation segments and the density and elongated nature of <001> loops and the various dislocation debris provide further information regarding the dislocation slip processes. Figure 4.15 (a) and (b) exhibit higher magnification images of a zigzag dislocation segment present in Figure 4.12. The b and d segments lie in the (110) slip plane approximately along <111> directions and therefore, have mixed character, while both the a and c components have line directions close to [010] and lie in the (100) plane. This means that a and c are both edge jogs which can only slip on the (100) plane. The Schmid factor for the [001](100) segments is about 0.35 at the [557] orientation and, therefore, these jogs should experience a lower shear stress. These features are illustrated schematically in Figure 4.15 (c). It is also interesting to notice that the primary slip plane segments b and d have a straight <111> line direction which is not common among those dislocations which do not have cross slip segments. However, according to Figure 4.13, the <001> dislocations on ( 110} planes have the lowest line energies in directions close to [111] and [1I T], in good agreement with the directions of b and d. This indicates that when the dislocations are pinned by closely spaced cross slip segments, they tend to align themselves in their lowest energy directions. This tendency is not as easily observed when the dislocation lines are pinned by cross slip segments that are separated by a distance larger than a critical value, since these dislocation segments will tend to bow out under the stress field.

A high density of small prismatic [001] loops are observed in the deformed NiAl as shown in Figure 4.12a. An interesting feature is that most of the loops are elongated along directions close to [011] which should indicate the mechanism for their formation and the









nature of the slip process. Since the densities of grown-in dislocations or annealing loops are extremely low in the NiAl crystals used in this study, as shown in Figure 4.11, there is no doubt that these prismatic loops are produced by plastic deformation, and that these loops are presumably formed by some sort of drag mechanism. Indeed, TEM images from deformed specimens revealed the presence of pinched-off loops from the trailing end of edge dipole segments (Figure 4.16). Such features could be produced by the pinning of screw dislocations by sessile jogs according to the Johnston and Gilman model. It is assumed that the [001] dislocations readily cross slip upon encountering obstacles as it has been suggested that cross slip is profuse in "soft" orientation NiAI (2, 8, 27). Thus, high densities of jogs could form via double cross slip. The mechanism proposed by Tetelman could also result in the observed prismatic loops elongated along [110] although it requires more restrained situations and may be only responsible for the dipoles or loops at the initial stage of deformation or at the regions between slip bands where cross slip of dislocations is less common. Based on the above discussion, the formation of [001] prismatic loops in [557] oriented NiAl single crystals can be explained based on the double cross slip mechanism as illustrated in Figure 4.16(c). After the dipoles and jogs are formed via this double cross slip process, the pinching-off of the prismatic loops will proceed by the Johnston and Gilman mechanism (Figure 2.12) if the heights of the jogs are appropriate.

After having considered the mechanisms of prismatic loop formation, the differences between the densities of loops in and out of slip bands may now be explained as the different scale of cross slip activities in and out of the slip bands. This result also indicates that cross slip is relatively easy in NiAL. However, in high purity NiAl, cross slip may only take place when dislocations encounter the stress fields imposed by other dislocations. This also indicates that very few other obstacles exist for dislocation movement in the high purity NiAI single crystals used in this study.









In order to study the dislocation activities outside the tangling region and investigate the characteristics of the dislocation bands, TEM foils perpendicular to the (110) slip plane and parallel to (001) were also made and investigated. Although only residual contrast exists in the TEM diffraction contrast images when the specimen is viewed with the Burgers vector [001] parallel to electron beam direction, as in the case of Figure 4.17, it is still obvious that the dislocation bands lie along [110] directions in these specimens, which verifies the [001](110) slip system. It is also noted that the residual contrast is strongest when g= 110 is used to form the image; this is due to the fact that the average direction of the <001> dislocations is close to the edge direction (i.e., u=[l10]), and for edge dislocations to have total invisibility, both g.b=0 and g.bxu=0 must be satisfied. As shown in Figure 4.17, using g=110 resulted in the weakest residual contrast, which is consistent with the edge line direction of [001] dislocations on (110) slip plane, i.e., u=[l 10].

In Figure 4.17, it can also be seen that the projection of dislocation segments which are not in band regions are usually straight and lie along [110]. This implies that these dislocations are strictly confined to the slip plane. However, the dislocation bands are broadening or bowing out from the [110] direction in many places. This suggests either intensive double cross slip or activation of new dislocation sources in the neighboring (110) planes of the already deformed layer. The latter is less likely because such activation of new dislocation sources requires the existence of high stress concentrations which is not in agreement with the low WHR (Table 4.1).

It has been suggested that cross slip is responsible for the broadening of slip bands or dislocation bands in many other materials, such as Fe-Si single crystals (107), copper alloys (108), neutron-irradiated Cu (109), and even h.c.p. crystals where cross slip is less frequently observed due to the splitting of dislocations into partials (110). In the case of [557] NiAI, the slip is planar due to its single slip nature. However, massive micro-scale









cross slip occurs in this material as indicated by the work of Ebrahimi et al. (26), Takasugi et al. (25) and Ball and Smallman (2). The distribution of prismatic loops in the deformed [557] specimens also indicates that massive cross slip occurred in or near the dislocation bands. Therefore, such broadening of the dislocation bands in NiAl is likely to be caused by double cross slip of [0011 dislocations. The process is essentially the same as that illustrated in Figure 4.16 for formation of the prismatic loops and jogs, and is described in further detail below.

At first, the screw component of a dislocation at the edge of a dislocation band cross slips under the stress field of the high density of dislocations inside the band. After moving on the cross slip plane for a certain distance h, which is a probability parameter depending on the local stress and temperature, the dislocation cross slips back onto a neighboring primary slip plane which is relatively free of dislocations. This will create a jog of height h, as illustrated in Figure 4.16. If h = 0.25 Gb
21r(I - v)(a - -1)

where G is the shear modulus, v Poisson's ratio and of the lattice friction, this cross slip either results in an edge dipole or a small jog which emits point defects or small debris during its subsequent motion. If h > h, a Frank-Read source is generated and, during continued straining, will produce loops and dislocations gliding on this new plane until further obstacles are met. In general, the dislocations inside the slip bands are pinned by numerous jogs as well as the high density of debris and loops and, therefore, are less mobile. The continued deformation of the specimen is mainly carried out by the new sources created at the edge of the dislocation bands and these bands become broadened as the amount of deformation increases.









Due to the formation and broadening mechanisms of the dislocation bands, their character and the overall substructure may reveal the role of cross slip in this material which may be important to understanding the mechanical behavior of NiAI. In the following sections, the effects of thermal treatments and deformation temperature on the dislocation substructure will be investigated in light of this understanding.

The dislocation substructure in more highly deformed [557] crystals (-10%) are shown in Figure 4.18. It can be seen that essentially all the dislocations are still the [001] type consistent with the stage I deformation exhibited by these crystals prior to fracture. The substructure consists of numerous discrete bands which tend to be thicker than those after 3% deformation, although the distance between the bands remains about same. Figure 4.18b shows that the trace of these dislocation bands lies approximately parallel to the Face of tensile specimens, which indicates these bands are still parallel to the (1 10) slip plane. This distribution is consistent with the development of the dislocation bands observed in the specimen deformed 3% and it indicates that the slip in these specimens is localized in the form of discontinuous dislocation bands lying parallel to the slip plane as illustrated in Figure 4.19. This can be explained by the broadening of the dislocation bands discussed above. These bands appear to develop from small scale dislocation tangling which initially could be caused either by impurity obstacles in the material or by interaction between opposite dislocations slipping on neighboring slip on planes.


4.2.2 [123] crystals

The dislocation substructure in a [123]-oriented specimen after 3% deformation is shown in Figure 4.20. It can be seen that dislocation bands similar to those developed in [557] specimens are also formed in [123] specimens. Since it was reported that the surface slip trace indicated {hk0) slip plane for [123] oriented NiAl single crystals (25), special attention was made to determine the slip plane of the dislocations. The results of g.b and









trace analyses verified that the majority of the dislocations have b=[001], and the dislocation bands lie along u=[1Il10]. Based on the knowledge gained from the dislocation substructure in [557] oriented specimens, we can expect a [001](1 10) slip system operating in this specimen. Other characteristic features in the [557] specimen are also present such as the (110) plane (primary slip plane) view which shows debris and loops elongated or stretched along [110] (Figure 4.20a). A BF TEM image taken with the beam direction near [001] shows the slip plane edge-on (Figure 4.20c), as can be seen, some dislocations are bowed out of their slip plane suggesting cross slip activity. However, the dominating dislocation line direction and the edge-on trace of the (110) slip plane unambiguously indicate that the primary slip plane is still (1 10), rather than the maximum resolved shear stress (MRSS) plane which should have an edge-on trace - 20� away from that of (1 10) in the B-[001] images. Nevertheless, a surface slip trace of the MRSS plane could be produced if massive scale composite slip on the primary (110) and secondary (010) slip plane occurs; this is more likely in the near-surface regions due to the complex stress state there.


4.2.3 [0111 Crystals


The dislocation substructure in [0111-oriented NiAl resembles that in the specimens oriented for single slip in that it is inhomogeneous in nature. Dislocations in [0111-oriented specimens exhibit a patch-like distribution when TEM foils are cut parallel to one of its slip planes, as shown in Figure 4.21. The slip vectors have been verified to be [001] and [010]. When viewing with the beam direction parallel to (010), as shown in Figure 4.21a and b, only [001] dislocations are visible. However, from the residual contrast, the projections of [010] dislocations can also be seen and appear to be tangled with the [001] dislocations, which suggests frequent interaction between the two slip systems. The [001] dislocation bands tend to lie along [100], as does the high density of prismatic loops which are located




























Figure 4.11 TEM BF image showing the typical microstructure of undeformed crystals.


Figure 4.12 Dislocations in a [557]S(1 10) specimens after 3% tensile strain at a strain rate of 10-4 S-: (a) g=002 near B=[I 10], (b) g=01 I near B=[100I, and (c) g =020 near B=[ 100] showing invisibility condition. The foil was cut parallel to (1 10) slip plane.





















































Figure 4.12 -- continued





















I


L(O)=1/cos(O) b 4L(O)=I


(a)


0 45 90 135 screw


180
screw


(b)
Figure 4.13 (a) Illustration of change in dislocation length after the relaxation (dashed line) and (b) Inverse of elastic energy E(O) of <001> dislocations on { 1101 planes in NiAI (Data compilation from (27)). 0 is the angle between the dislocation line direction and its Burgers vector. 0=90' corresponds to the edge orientation while 0=0 or 180 corresponding to screw orientations.


1/E(O)




















Jb




Figure 4.14 Schematic illustration of the forming of the bowed-out 40-50� dislocations and sharp tip. The dislocation will bow out in the edge direction due to higher mobility of the screw segments. The sharp tip could be due to relaxation of the short screw segment.

















(a) 0(b)

Figure 4.15 Higher magnification images of the zigzag dislocations in (a) Figure 4.12a, g=002 near B=[1 10], and (b) Figure 4.12b, g=0 11 near B=[I00]. (c) Illustration of the zigzag segments.













I ~ ~[01O] ?I
I __


I --
~I/ [111]
I

I


or I
(c)


(100)


~/




I


l [001] / II
" W[11 0] /I I/
- 111o)--


Figure 4.15 -- continued


Figure 4.16 TEM images showing pinched-off of prismatic loops (a and b) and (c) a schematic illustration of the formation mechanism.



































Cross slip plane


(001) /


(10) (001 /

top


I,/
bottom /-L / \\ _/ /, /


Figure 4.16 -- continued



















































Figure 4.17 Dislocations in a [557] S(1 10) specimen. 3% strain at a strain rate of 10-4 s'. Foil face parallel to (001): (a) g=1 10, strong residual when g.bxu * 0, (b) g=020, g.bxu * 0, (c) g= 110, weakest residual contrast when g.bxu = 0 and (d) g=200, g.bxu # 0. Images taken using a similar deviation parameter s.























































Figure 4.17 -- continued




















































Figure 4.18 Dislocation substructures in a [557]S(1 10) specimen after 10.3% deformation. Foil normal about 15' from (T10) slip plane normal. BFTEM images showing: (a) dislocation bands, (b) the dislocation tangling with B near (1 10), and (c) the invisibility condition.































Figure 4.18 -- continued


(110)


Figure 4.19 Schematic illustration of the distribution of dislocation bands in [557] oriented NiAl single crystals deformed at RT. (110) is the slip plane. It can be seen that the slip is localized in the form of patch-like dislocation bands.







65






















(a)



















(b)

Figure 4.20 Dislocations in [123] oriented specimen after 3% deformation: (a) (110) slip plane view, (b) (I 11) plane view and (c) (001) plane view.





























Figure 4.20 -- continued


Figure 4.21 Dislocations in [011] oriented NiAl after 6.1% deformation at RT showing:
(a) dislocation bands along [100], (b) prismatic loops elongated along [100] and jogged dislocations , and (c) dislocation bands when B near [011 .

























































Figure 4.21 -- continued






















Cross slip plane


(010) [001] /I/ 'I,


/,7


I /
/
/
'/-------------


/\
/ 4'


I',


v / Ip e
/ Edge dipole


/ I /'
------_k /


Figure 4.22 Illustration of proposed double cross slip mechanism for forming dipoles and pinched-off loops in [011] oriented NiA1 single crystals.









primarily within the bands. Dislocations are often severely jogged and zigzagged, as shown in Figure 4.21b, which could be due to the pinning of either jogs or interacting dislocations from another system, or dislocation relaxation due to the higher energy of the screw dislocations. Since [100] is the edge direction of both [001](010) and [010](001) dislocations, the characteristics of the dislocation substructure are similar to those observed in [557] oriented crystals. Thus, it is clear that similar dislocation behavior exists in both single and double slip oriented NiAI crystals. Based on the similar analysis that was used to determine the slip plane of [557] and [123] crystals, the slip systems in [011] crystals can be determined to be [001](010) and [010](001). Since [001] and [010] dislocations always tangle together, as shown in Figure 4.21b, the interaction between these two systems may be the origin of dislocation tangling and bands. However, the high density of prismatic loops indicates that cross-slip is common, as it is unlikely that such jogs are produced via interaction between the two kinds of dislocations. Similar to the dislocation substructure in the single slip specimens, these dislocation features can be explained using a similar double cross slip mechanism (Figure 4.22) as the one for the single slip orientations; this predicts the formation of pinched-off loops elongated along [100] from the edge dipoles created via double cross slip.

Therefore, in [0111-oriented crystals, both the [010](001) and [001](010) slip systems are operating. The interaction between [001] and [010] dislocation is responsible for the higher work hardening rate (Figure 4.1.1) and may create initial dislocation tangles which subsequently broaden via cross slip of both kinds of dislocations.


4.3 The Effect of Heat Treatment on Slip Behavior at Room Temperature

Two specimens from the fractured furnace cooled (FC) and water quenched (WQ) [557] S(110) NiAI tensile specimens were chosen for dislocation substructure analysis.










The TEM foils were cut parallel to the Face plane. The WQ specimen was deformed to about 2% plastic elongation before fracture and the FC specimen -8%.

The dislocation substructure in the FC [557] specimen consisted of bands (Figure 4.23) similar to those observed in the [123] and [557] specimens described in Section 4.2, resulting from the double cross slip of dislocations on (110) planes. BFTEM images taken with a beam direction near [001] clearly show the images of the dislocation bands edge-on (Figure 4.23a). It can be seen that these bands are broadening in some regions although some bands exhibit a relatively straight edge. As already discussed in Section 4.2, this kind of broadening is caused by intensive double cross slip from the heavily deformed primary slip plane to the relatively dislocation-free neighboring primary slip planes. It is important to note that there is very little slip activity in the regions between slip bands (Figure 4.23a).

A slip-plane view of the dislocations was also obtained and the following features are obvious (Figure 4.23c): (1) most dislocations are in the band regions and (2) a high density of debris and prismatic [001] loops elongated along [110] inside these dislocation patches (arrows). These features are similar to what was observed in the [557] oriented specimens discussed in Section 4.1. and, if there is any difference, it is that there is an even lower density of dislocations in the regions between the slip bands. Therefore, it appears that furnace cooling did not affect the slip behavior in the [557] oriented specimens as compared with the homogenized specimens studied in Section 4.1, except that there are even fewer obstacles for dislocation glide in FC crystals as indicated by the dislocation-free regions between the dislocation patches. The higher density of debris in the bands and the thicker bands indicate that the deformation occurred by localized slip, i.e., by the cross slip of dislocations.

In the WQ specimen however, a radically different dislocation structure was observed as shown in Figure 4.24a. It is obvious that the dislocation structure is much more uniform compared with that in the FC specimen. Although there is still slip









localization as indicated by the dislocation tangling, debris and small loops distributed throughout the regions between the bands (Figure 4.24c-f), the structure suggests more uniform slip processes than in the FC specimen. Since it has been shown that the primary mechanism for forming debris and small prismatic loops in NiAI is by double cross slip, the dislocation substructure in Figure 4.24 indicates that cross slip was occurring everywhere, i.e., outside as well as within the bands; this indicates a more interrupted dislocation movement in the WQ crystals.

It can be seen that the loops and most of the dislocations are in contrast with g= 101 and g=002, and out of contrast with g=1 10 and g=200. Consequently, the Burgers vector of the majority of the dislocations and loops is still [0011. However, unlike in the FC crystals, there are a small fraction of dislocations belonging to a secondary vector in the WQ crystals, as shown in Figure 4.24b. g.b analysis shows that these dislocations are either [100] or [010] dislocations.

Tilting the TEM specimen to the [001] zone places the slip plane edge-on as shown in Figure 4.24d. Once again, the numerous loops spread out over the region and it again shows that slip in WQ specimens is less localized than in FC specimens and other single slip oriented specimens studied in previous sections. Besides showing some traces of primary dislocation bands in the [110] direction, which indicates the primary slip system is [001](110), the edge-on image in Figure 4.24g also shows straight traces of [001] dislocations lying perpendicular to the (110) slip plane (B2 in the figure) and traces of prismatic [001] loops along both the [110] (LI in the figure) and [010] (L2 in the figure) directions. Except for the LI loops, other features were not observed in the slower cooled specimens. The double cross slip mechanism of [001](110) dislocations proposed for the formation of [001] prismatic loops elongated along [110] directions can not be used to explain the loops elongated along [010] in the [001] projection. Illustrated in Figure 4.25, one possible mechanism of forming these traces is the slip by a <001 > {010) system which









could leave debris or prismatic loops elongated along { 100 } directions via double cross slip. The straight projection trace of [001] dislocations along the trace of [110] planes can only be explained by slip on [1101 planes, which is not understandable at this time as there is no resolved shear stress for the [001 ](110) slip system. Nevertheless, all these features suggest that cross slip in NiAI is common and not only confined to the dislocation band region as in the case of FC or homogenized [557] crystals. Apparently, the water quench changed the micro-scale slip process of [557] oriented NiAl single crystals. This may represent changes in the dislocation mobility and cross slip behavior in the vacancy hardened NiAl. More work will be needed to address these issues further. The difference in the amount of the plastic deformation (8% vs. 2%) that the FC and WQ specimens experienced before fracture may also account for some of the differences in the dislocation substructure. However, it is unlikely that these differences in the prismatic loops and the occurrence of secondary slip are due to the difference in the amount of deformation, since these prismatic loops and the secondary dislocations should not disappear at RT once they are produced.

It is, however, possible to explain some of the changes in the mechanical behavior of the WQ NiAl single crystals with the dislocation structures revealed here. Compared with slowly cooled specimens where the glide of [001] dislocations is easy at the beginning of deformation and where no cross slip will take place in the earliest stages, it appears that the [001] dislocations in the WQ specimens have difficulty gliding from the beginning and that cross slip is induced prior to any interaction with the stress fields of other dislocations. The fact that the prismatic loops are no longer only elongated in one direction suggests that slip may not be as confined to the primary (110) slip plane. The lesser tendency to form dislocation bands (or localized slip) could also be attributed to the difficulty in the movement of cross-slipped dislocations as was shown in a previous section where the broadening of the bands is assisted by the cross slip of dislocations. Due to the difficulty





















































Figure 4.23 The dislocations in the FC [557] oriented NiA1 single crystals tested to 8% at a strain rate of 10-4 S-: (a) g=200 and B near [001], (b) g=101 and B=[ 111], and (c) g=002 and B near [110]. Arrows point to elongated loops.































Figure 4.23 -- continued


Figure 4.24 TEM BF image showing dislocations in the WQ [557] oriented NiAl single crystals: (a) g=101 and B near [111], (b) g=l 10 and B near [11 1], (c) g=002 and B near [110], (d) g=200 and B near [001], residual contrast, (e) high magnification of (c), (f) g=002 and B near [1 10], (g) high magnification of (d).



















































Figure 4.24 -- continued























































Figure 4.24 -- continued












































I


Figure 4.24 -- continued












































Figure 4.25 The illustration of the possible mechanisms which can explain the dislocation features shown in Figure 4.24g: double cross slip of [001](110) dislocations producing (001] prismatic loops L1; double cross slip of [001 ](010) dislocations resulting in L2 loops; [001 ] dislocations slip on (110) plane resulting trace B2; and B I is the primary [001](110) dislocation bands.









in the motion of cross slipped dislocations, the stress concentrations may develop to such a high value at dislocation tangles that even secondary dislocations ([010] and [100]) are generated (Figure 4.24b and d). It is also apparent that, in the WQ specimens, the stress concentration will develop more rapidly at localized strain points with much smaller strain than in the slowly cooled specimens, as also indicated by the difference between the work hardening rates of the WQ and FC specimens (Table 4.2). Hence, the room temperature tensile plasticity is greatly reduced. This behavior could be readily attributed to the dislocation-vacancy interactions caused by supersaturated thermal vacancies in the WQ specimens, consistent with the much higher CRSS for the WQ specimens (Table 4.2), although dislocation pinning by vacancy clusters is also possible as indicated by the shear loops in the slip plane (Figure 4.24f). However, the detailed mechanisms still need to be investigated.


4.4 Effects of Temperature and Strain Rate on the Slip Behavior of Soft-Oriented NiAl


Dislocations in tensile specimens oriented along [557] and tested at either 120K or 77K are shown in Figure 4.26. As shown in Figure 4.26a-c, standard g.b analysis indicated that the dislocations have the same [001] Burgers vector as in the RT deformed [557] crystals. However, unlike the dislocations in the [557] specimens deformed at RT, the dislocations after sub-ambient deformation tend to be long and relatively straight, especially in the specimens tested at 77K (Figure 4.26e). The projected images of the dislocations are consistent with a true line direction uT=[ 1101; this is consistent with a [001](110) slip system and indicates edge or near-edge character for the majority of the dislocations which suggests an even lower mobility for the <001> edge components. The dislocation substructure in the specimen deformed at 120K displays similar features such as zigzag dislocation lines, edge dipoles and prismatic loops elongated along [110] (Figure 4.26d). These features suggest that the cross slip of dislocations is also taking place at









these lower temperatures. However, the absence of dislocation patches and the lower density of loops suggest that the cross slip activity at 120K is not as intense as at RT. This kind of dislocation substructure is consistent with the Arrhenius representation of the yield stress (Figure 4.4) which indicated no change in the slip behavior at temperatures below RT.

Since the Arrhenius representation of the yield stress suggests a change in slip behavior at temperatures above 500K, and since it has been reported that there is no change in the slip systems in soft-oriented NiAl single crystals, it is of interest to know what kind of changes occurred in the dislocation substructure after deformation at these higher temperatures.

Figure 4.27 shows the dislocation substructure in a [557] crystal deformed to about 18% deformation at 473K using different strain rates. It can be seen that, in general, the dislocation distribution resembles that in RT deformed [557] specimens, which indicates no substantial difference between RT and 473K slip behavior. However, it appears that the strain rate does influence the dislocation substructure. The dislocations in the specimen tested at the higher strain rate appear more stretched out and lying closer to the edge orientation which suggests either the difference between the mobility of screw and edge dislocations became larger, or the screw dislocations were dragged more severely by the sessile jogs. Since the mobility of unpinned or unjogged dislocations is unlikely to be affected by strain rate, it is more probable that the above feature is due to more severe jog dragging with increasing velocity of the screw dislocation. It can also be seen that the stress concentration at the dislocation bands in the higher strain rate specimen appears to be larger since the activation of secondary slip vectors occurs (Figure 4.27b) in contrast to the behavior in the lower strain rate specimen (Figure 4.27d). This may be due to the fact that the stress concentration at the front of the slip bands could not be readily released by cross slip or by generation of new dislocation sources via cross slip, since the movement of the







81


jogged screw dislocations and the cross slip activity can not keep up with the higher imposed strain rate. As a result, the specimens tested at the higher strain rate tend to have lower tensile elongations at this temperature as shown in Section 4.1.

The scenario changes somewhat at even higher temperatures (e.g., 673K) where the specimens can be deformed easily to very large elongations. Figure 4.28 exhibits the dislocation substructure in a [557]S(1 10) specimen deformed at 673K. Since at this temperature the tensile specimens form necks, the samples cut from the necked segments of the tensile specimen should have undergone higher local strains than the other parts of the specimen. It can be seen that there are some features in the dislocation substructure resembling the dislocation features in RT deformed specimens such as the high density of loops and debris (Figure 4.28a) and jogs pinning the dislocations (Figure 4.28c). However, the existence of dislocation bands is less obvious and the deformation is more uniform than that at RT and 473K. Furthermore, some sort of sub-boundary has already started to form at this temperature (Figure 4.28b) suggesting that the diffusion rate has reached a level that allows climb to assist in the recovery process. The recovery of the dislocation substructure also is indicated by the relaxed dislocation lines and the larger average size of the prismatic loops, as shown in Figure 4.28c.

The dislocation substructures in Figure 4.28 should represent the slip in the later stages of the deformation where necking causes inhomogeneous deformation throughout the gage of the specimen and, for those dislocation substructures formed during the earlier stage of deformation, there was sufficient time to allow recovery of the structure. To investigate the slip behavior in the earlier stage of deformation, a [123] crystal was deformed to 3% at 723K before the TEM samples were cut from it. Figure 4.29 shows the dislocation substructures in this specimen. It is obvious that, at the earlier stage of deformation, both slip localization and formation of dislocation bands occur. However, the bands are much more broadened and much less dense than those usually observed at RT.









The higher magnification image of the dislocation bands in Figure 4.29b clearly shows that extensive cross slip occurred and the traces of most of the cross slip segments lie along the [200] direction (arrows), which suggests a (010) cross slip plane. The above results indicate that although the slip behavior at these intermediate temperatures still remains the same (i.e., the slip is still localized and the slip system is the same), the local strain and stress concentration can be released more effectively probably because of easier cross slip, higher mobility of the jogged screw dislocations (due to faster short-range diffusion), and faster recovery of the dislocation substructures due to climb.

Once again, by testing at even higher temperatures, the only active slip vectors in soft-oriented NiAl single crystals are <001> type. The slip vectors other than [0011 are present presumably due to the existence of necking in the specimen gage which changes the local stress state. The highest elongation could be achieved at certain temperatures depending on the strain rate and orientation, due to the balance between strain hardening and recovery (6, 112). Figure 4.30 shows the dislocation substructures in the crystals deformed at 1073K at strain rates of 10' s-' and 102s-'. It can be seen that the deformation at elevated temperatures in soft NiAl crystals is controlled by both the glide and climb of cube dislocations. In the case of the slower strain rate, large dislocation networks and subboundaries form in the specimen (Figure 4.30a-b). These dislocation networks often consist of only [001] dislocations (Figure 4.30a) indicating the occurrence of a considerable amount of climb. In the case of the faster strain rate, the dislocation networks are comprised of two types of <001> dislocations interlaced together (Figure 4.30c); this indicates that the contribution from climb decreases at the higher strain rate as expected. The tensile elongations of the corresponding specimens can also be explained with the dislocation behavior shown in the figure. Since the amount of recovery in the specimen deformed at a strain rate of 10.2s-' was lower, the specimen exhibited better resistance to




















































Figure 4.26 Dislocations in [557] oriented NiAl deformed -1.3% at 120K (a-d): (a), g=002 near B=[ 100], (b) g=01T near B=[100], (c) g=020 near B=[ 100] showing dislocation out of contrast and (d) g=002 near B=[ 100] showing prismatic loops; and deformed at 77K: (e) long straight dislocations, g=O1 I near B=[100]. Strain rate = 104s-.





















































Figure 4.26 -- continued











































Figure 4.26 -- continued





















































Figure 4.27 BFTEM images showing dislocation substructure in the [557] specimens deformed at 473K at a strain rate of 10.' s' (a, b) (elongation=18%) and 10 s' (c, d) (elongation=19%): (a) slip plane view; (b) slip plane edge-on; (c) slip plane view and (d) slip plane edge-on.




























(c)


Figure 4.27 -- continued


















































Figure 4.28 Dislocations in [557]S(110) specimen tested at 673K: (a) dislocation substructure, (b) subboundary, and (c) dragging of the dislocations by jogs. TEM foil cut parallel to the Face from the necked region in the gage of tensile specimen. Strain rate = 10s . g=O11 near B=[II1].








































Figure 4.28 -- continued


















































Figure 4.29 Dislocation substructure in the [123] oriented specimens after 3% deformation (strain rate = 104 s-) at 723K showing (a) bands and (b) cross slip segments out of the (I 10) primary slip plane (arrow points). B near [001].



















































Figure 4.30 Dislocation sub-boundary and networks in the [557] oriented specimens deformed at 1073K and a strain rate of (a, b) 10. s', and (c) 10-2 S-.
















~1






0
0 c~.














0









necking and higher tensile elongation than that tested at the lower strain rate (10-4 s-) (Figure 4.5).

In general, the temperature affects the deformation of soft-oriented NiAl single crystals via the gradual increase in the activity of both cross slip and climb mechanisms with increasing temperature. Likewise, the strain rate affects deformation in a similar manner and the combination of temperature and strain rate dictates the relative WHR and recovery rates of NiAl single crystals. No fundamental change in the slip behavio, is observed over the temperature range studied.


4.5 Slip Behavior of Hard-Oriented NiAI Single Crystals


In order to identify the slip behavior in <100>-oriented NiAI and NiAI-Si single crystals, TEM foils were cut from tensile specimens after 2-5% tensile deformation at 473K, 523K, 573K, 723K, and 873K. The foils were cut parallel to the (100) (NiAI) and (110) (NiAI-Si) planes from specimens whose tensile axis was [0011.


4.5.1 Dislocations in the specimens tested below 573K


Only cube <001> dislocations were observed in specimens deformed below 573K. The deformation appears quite heterogeneous since in some TEM foils, no dislocations were observed while in others, dislocation substructures such as that shown in Figure 4.31 were observed. This particular micrograph was taken from a specimen tested at 523K to

-2% elongation before fracture. Standard g.b and trace analyses indicated that these dislocations are all [010] dislocations mostly lying on the (100) plane. There are many prismatic loops elongated along [0011, similar to what is observed in the [011] oriented RT specimens. However, since there is no resolved shear stress for <001> slip when the specimen is oriented accurately along <100> directions, activation of cube slip is not expected. One possible mechanism of producing cube slip in <100> oriented NiAI is by




Full Text

PAGE 1

A STUDY OF LOW AND INTERMEDIATE TEMPERATURE SLIP BEHAVIOR OF HIGH PURITY NiAl SINGLE CRYSTALS By JIAN HU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS OF THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1997

PAGE 2

TO MY PARENTS

PAGE 3

ACKNOWLEDGMENTS Pursuing a Ph.D degree at an age of over 30 years old and in a totally new culture and language environment presented an exciting challenge for me and this challenge could not have been overcome if I had not received the kind and invaluable guidance and advice from all of my supervisory committee members: Dr. M.J. Kaufman, Dr. F. Ebrahimi, Dr. R.T. Dehoff, Dr. R. Reed-Hill and Dr. A. Kumar. I am especially thankful to Dr. M.J. Kaufman, my committee chairman, for his tireless technical advice and instruction, as well as for his constant encouragement to me for becoming a better person. I can never thank him enough for all his help to my growth in my career, and to my growth in other aspects of my life. I would also like to extend my sincerest thanks to Dr. V. Levit for his countless help during the years this research project has been going on. Dr. V. Levit and his wife have been very generous in providing me with excellent experimental materials and help in conducting experiments, as well as invaluable advice and suggestions throughout this study. Without their help, the completion of this research would not have been possible. I am also very grateful to other graduate students and colleagues, especially Zheng Chen, Sanjay Shrivastava, Mark Weaver, Andy Duncan and Yongjin Lim, who constantly offered valuable suggestions and help to me. Finally, I could never have come this far without the love and encouragement from my parents and my brothers. I am deeply indebted to them and hope this dissertation expresses my heartfelt thanks to them. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS iii ABSTRACT vi CHAPTER 1 INTRODUCTION 1 Background 1 Approach 4 2 LITERATURE REVIEW 6 2.1 Physical Properties of NiAl 6 2.2 Yield Behavior of NiAl 7 2.3 Slip Behavior 8 2.3.1 Theoretical Calculations 9 2.3.2 Observed Slip Systems and Dislocation Behavior 1 1 2.3.3 Effect of Heat Treatment, Alloying Elements and Impurities 12 2.4 Ductility and BDTT 13 2.4.1 Ductility of Single and Polycrystalline NiAl 14 2.4.2 Brittle-to-Ductile Transitions (BDT) in NiAl 15 2.5 Formation of Dislocation Dipoles and Loops During Plastic Deformation 18 3 EXPERIMENTAL PROCEDURE 27 3. 1 Sample Preparation 27 3.1.1 Single Crystal Growth 27 3.1.2 Orientation Determination 28 3.1.3 Heat Treatment 28 3. 1 .4 Preparation of Tensile Specimens 28 3.1.5 Preparation of Transmission Electron Microscopy Specimens 29 iv

PAGE 5

3.2 Mechanical Testing 30 3.3 TEM Dislocation Analysis and SEM analysis 31 4 RESULTS AND DISCUSSION 33 4.1 Mechanical Properties 33 4. 1 . 1 Soft Orientation at Room Temperature 33 4.1.2 Soft Orientation at Non-ambient Temperatures 36 4. 1 .3 Heat Treatment Effects 37 4. 1.4 Hard Orientation 38 4.2 Slip Behavior of Non-<001> Oriented High Purity NiAl Single Crystals at RT 47 4.2.1 [557] Crystals 47 4.2.2 [123] crystals 53 4.2.3 [Oil] Crystals 54 4.3 The Effect of Heat Treatment on Slip Behavior at Room Temperature 69 4.4 Effects of Temperature and Strain Rate on the Slip Behavior of Soft-Oriented NiA179 4.5 Slip Behavior of Hard-Oriented NiAl Single Crystals 93 4.5.1 Dislocations in the Specimens Tested Below 573K 93 4.5.2 Slip Behavior at 573K 94 4.5.3 Slip Behavior at Temperatures Above 573K 98 4.6 Slip Traces and Fracture Surfaces 1 12 4.6.1 Soft Orientation 112 4.6.2 Hard Orientation 1 14 4.7 General Discussion 121 4.7. 1 Deformation of NiAl Single Crystals 121 4.7.2 Tensile Ductility of Soft Orientation 123 4.7.3 Tensile Ductility of Hard Orientation NiAl Crystals 129 4.7.4 Potential of NiAl Single Crystals as High Temperature Structural Material.... 132 5 SUMMARY AND CONCLUSIONS 136 LIST OF REFERENCES 140 BIOGRAPHICAL SKETCH 146 v

PAGE 6

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A STUDY OF LOW AND INTERMEDIATE TEMPERATURE SLIP BEHAVIOR OF HIGH PURITY NiAl SINGLE CRYSTALS By JIAN HU December, 1997 Chairman: Dr. Michael J. Kaufman Major Department: Materials Science and Engineering The slip and mechanical behavior of high purity NiAl single crystals at temperatures between 77 and 1073K has been investigated in an effort to understand their tensile properties along both non-<100> "soft" and <100> "hard" orientations and to obtain a better understanding of the observed sensitivity of the behavior to impurity elements and thermal history. It is shown that, at a strain rate of 10" 4 s'\ the high purity NiAl single crystals exhibit over 30% tensile elongation at RT when tested along certain soft orientations and the tensile elongation increases gradually from 77K (-0%) to 473K (50%) before increasing more rapidly due to the onset of dynamic recovery. The cross slip behavior of NiAl was also studied and it is shown that, in softoriented NiAl crystals, cross slip occurs readily at room temperatures and that slip is very inhomogeneous as determined from the presence of discrete dislocation bands. A double cross slip mechanism is proposed to account for the observed high density of debris and small prismatic loops in the soft-oriented NiAl crystals after deformation and the vi

PAGE 7

broadening of the dislocation bands. It is also proposed that the tensile ductility of soft NiAl crystals is controlled by the rate of cross slip and local short range diffusion which affects the development of dislocation bands and thus the local stress concentrations that ultimately lead to brittle fracture. The slip behavior in the hard NiAl single crystals is also investigated. It is concluded that the sudden increase in tensile ductility of the crystals in this orientation is due to the activation of the <011>{011} slip systems. The results of this study also indicate that, in high purity stoichiometric NiAl, <011> dislocations dominate the deformation near 573K but become unstable at and above 723K. The <01 1> dislocations also do not cross slip at 573K and contradictory to the predictions of theoretical calculations, these dislocations exhibit a strong tendency to lie along <1 1 1> directions and assume a skew configuration in the slip planes. The slip behavior of hard orientation NiAl0.3Si single crystals was also studied and the results indicate that Si does not strongly affect the slip behavior of <100> oriented NiAl yet increases the tensile elongation at and above 673K by some unknown mechanisms. It is proposed that the thermally activated kink motion of <011> dislocations along their <111> line directions accounts for the BDTT of <001> oriented NiAl single crystals. vii

PAGE 8

CHAPTER 1 INTRODUCTION Background In the past few decades, the potential application of intermetallic compounds in jet engines revived the worldwide interest in these materials due to the need to increase the operation temperatures of jet engines for better efficiency and the limited melting temperatures of the currently used superalloys. Many intermetallic compounds such as aluminides, silicides, carbides, and other novel intermetallics have been tested and studied in the drive to evolve usable high temperature structural materials from them. NiAl is one of the most promising and most extensively studied among these intermetallic compounds because of (1) its low density (-5.85 g/cm 3 ), which could result in the reduction of engine weight, (2) its high melting point (191 1-1955K), which is -300K higher than that of nickel, (3) its large stability range, which could ease alloying and manufacturing difficulties, (4) its good oxidation resistance, (5) its excellent electrical and thermal conductivity and (6) its low cost. However, like many other intermetallic compounds, NiAl exhibits very limited ductility and toughness at room temperature which makes it prone to handling and impact damage. Its poor high temperature strength further hinders its application as a structural material although it finds limited industry applications in areas such as coatings on turbine blades (1). Although the yield stress of NiAl and its temperature dependence strongly depend on the orientation and various metallurgical factors, with remarkable difference between polycrystalline and single crystal forms, it normally shows characteristics resembling those 1

PAGE 9

2 of BCC metals. That is, it has a strong temperature dependence at temperatures below about 0.15 T m (about 300 K) and a relatively temperature independent region at intermediate to high temperatures (2, 3, 4, 5, 6, 7). However, the temperature dependence of the yield stress of NiAl single crystals is reported to be strongly orientation dependent and very sensitive to both the impurity elements and deviations from stoichiometry. In single crystal form, NiAl has a much higher yield stress when oriented along <100> (cube) directions at temperatures below 600K although the high temperature yield stress is relatively insensitive to crystal orientation (3, 8, 9). The ductility of NiAl is also strongly orientation dependent and extremely sensitive to many metallurgical parameters such as impurity concentration, and Ni/Al ratio, and is in general very sensitive to strain rate. It has been shown that polycrystalline NiAl has virtually no ductility at room temperature (RT) (10). For single crystal NiAl, no tensile ductility has been observed in <100> oriented (hard) NiAl below -589K (4) although limited compression plasticity is obtainable (7, 8). However, in non-<100> oriented (soft) NiAl single crystals, limited (typically <2%) tensile elongation has been obtained at temperatures at RT and below (3, 11). As temperature increases, it is reported that all forms of NiAl undergo an abrupt increase in ductility at the so-called brittle-to-ductile transition temperature (BDTT). The exact temperatures of the transitions have been found to depend on various metallurgical parameters in the case of polycrystalline NiAl as well as on the crystal orientation in the case of single crystals. This prompted many research efforts to understand the mechanisms for these transitions. However, except in the case of polycrystalline NiAl where the onset of local short circuit diffusion such as dislocation climb is generally believed to be the mechanism responsible for the abrupt increase in ductility (12), the actual mechanisms for both non-<100> and <100> orientation single crystal NiAl are still not clear.

PAGE 10

3 It is suggested that some of the most important factors controlling the low temperature ductility of NiAl include the concentration of interstitial impurities in the material and the thermal history of the material. Some recent results indicated that reducing the concentration of interstitial impurities either by purifying the material or doping with gettering elements typically decrease the room temperature yield stress and sometimes result in slightly better room temperature ductility (13, 14, 15, 16). Indeed, like in the conventional BCC metals, interstitial species such as C and O have been shown to be responsible for the strain aging phenomena apparent in lower purity NiAl as shown in the recent work of Weaver et al. (17, 18, 19, 20) and Hack et al. (21). However, it is still not clear if the poor low temperature ductility is an intrinsic characteristic of NiAl as there has been limited work concerning the deformation behavior in higher purity NiAl, i.e., where there is any change in the slip behavior. The slip vectors responsible for the deformation of NiAl single crystals are <001> slip when oriented along any of the soft non-<001> orientations (7, 8, 22, 23, 24). In most reports, the observed actual slip planes are {100} and { 1 10} depending on the orientation (i.e., Schmid factor). However, slip on higher index {hkO} planes has also been reported by Takasugi et al. (25) and by Ball and Smallman (24). The latter authors explained the apparent {hkO} slip as resulting from cross-slip between orthogonal { 1 10} planes. Pascoe and Newey also reported that, at very low temperature (77K), slip is confined to {110} planes regardless of the crystal orientation (7) although this view is not widely accepted. Various authors (26, 27) reported that the cross-slip of <001> dislocations is relatively easy at and above room temperature. However, detailed information regarding the crossslip behavior of dislocations in NiAl such as the cross-slip plane and its effects on the ductility and other mechanical properties of NiAl single crystals is still lacking and controversial.

PAGE 11

4 Because of the negligible resolved shear stress for cube slip, the slip systems in the NiAl single crystals oriented along <100> are different from those oriented along the soft directions. For the hard orientation, the results of previous studies (3, 7, 22, 27, 28, 29, 30) indicate that (1) below 600K, <1 1 1> slip dominates, (2) between 600K and 800K, there is a transition from <1 1 1> to non- slip (e.g., <1 10>) and (3) above 800K, deformation is by < 1 1 0> {110} slip and <001> climb. Unlike the case of the soft oriented crystals where there is no change in slip system associated with the BDTT, the sudden increase in tensile ductility of <100> crystals appears to correspond to a transition from <1 1 1> to non<111> slip; this occurs about 100K higher than that of the soft crystals. However, the correlation between this transition in slip behavior with the BDTT in hard NiAl crystals has not been determined, and the details of the slip behavior of <100> oriented NiAl single crystals around the temperature range where this transition occurs are still not understood, especially in stoichiometric NiAl. It is obvious that, despite a fair amount of investigation, many details of the slip behavior in NiAl single crystals have not been studied systematically. Questions such as the role of cross-slip, the development of dislocation substructure with strain, the effect of thermal treatment on the slip behavior, the relationship between slip behavior and ductility, and the details of the orientation dependence of slip behavior still remain largely unanswered. Clearly, without a clear understanding of the intrinsic slip behavior in NiAl, any approaches to improve its ductility and toughness are difficult to justify. Therefore, a systematic study of the slip behavior in high purity NiAl single crystals is necessary. Approach In this work, high purity NiAl single crystals will be tested along a few carefully chosen orientations in order to generate a reliable mechanical property database. An explanation of the mechanisms responsible for the mechanical behavior, especially for the

PAGE 12

5 change in the tensile ductility of NiAl single crystals with temperature and thermalmechanical history will be attempted by performing detailed investigation of the slip and fracture behavior after testing at various conditions (temperature, strain rate, orientation, etc.). It is the purpose of this study to (1) achieve a better understanding of the slip behavior of NiAl single crystals by systematic analyses of dislocations in high purity single crystal NiAl while such factors as the impurity level, doping elements, heat treatment and specimen geometry are controlled, and (2) correlate these observations with both the deformation and fracture behavior of NiAl single crystals as a function of orientation, strain rate and temperature.

PAGE 13

CHAPTER 2 LITERATURE REVIEW This review will summarize previous results concerning the deformation and slip hehavior of NiAl, as well as the factors affecting the deformation of NiAl in order to bring to light some important areas that need to be better understood. 2. 1 Physical Properties of NiAl NiAl is a Hume-Rothery (3-phase electron compound with a B2 ordered crystal structure as shown in Figure 2. 1 . The stoichiometric NiAl melts congruently somewhere in the range of 191 1-1955 K (31, 32). NiAl has a relatively wide single-phase composition range (from about 45 to 60 atomic percent Ni) as indicated by the Ni-Al binary phase diagram (Figure 2.2 (33)). Unlike many other B2 compounds which disorder before melting (e.g., FeAl), NiAl remains ordered to the melting temperature consistent with its large negative heat of formation (-72 KJ/mol (34, 35)). Quenching from the vapor at rate of ~10 8 K/sec does not suppress the ordering of NiAl (36) so it is believed that the ordered structure is stable up to its congruent melting temperature. The calculations and experimental measurements of the electronic band structure are in good agreement and it is generally concluded that a very strong atomic bond exists between Ni and Al along the <1 1 1> directions, consistent with the strong ordering of this material up to the melting temperature, and much weaker bonding along the <100> directions (37). NiAl is elastic anisotropic with an anisotropy factor (A=2C 44 /(C n -C l2 )) between 3.2 (38) and 3.3 (5). This factor is only weakly temperature dependent for near6

PAGE 14

7 stoichiometric compositions, and more strongly dependent on composition (38) with a trend of lower anisotropy for Al-rich and higher anisotropy for Ni-rich compositions (Figure 2.3). When the composition becomes Ni rich, an anomalous temperature dependence is observed, and this has been related to a softening of NiAl along <1 10>{ 1 10} (39, 40). Since the dislocation energies and core structures are affected by the elastic anisotropy, any change in the elastic anisotropy factor could influence the slip behavior. It is claimed that a transfer of electrons from the Al s band to the Ni d band occurs (41) and results in the strong Ni-Al bond energy. Consistent with such electron transfer, it is reported that off-stoichiometric binary NiAl will form constitutional vacancies (vacancies on Ni sites) when it is Al-rich and Ni anti-site point defects (Ni on Al sites) when it is Nirich, based on measurements of the composition dependence of the lattice constant and density of NiAl (42, 43) (Figure 2.4). These constitutional defects are reported to display short-range order to avoid having like species as nearest neighbors (44) and the existence of constitutional defects makes the properties of NiAl very sensitive to any deviation from the stoichiometric composition. 2.2 Yield Behavior of NiAl The flow stress of NiAl single crystals strongly depends on the orientation. When oriented along the "hard" <100> direction, NiAl has much higher flow stresses at low and intermediate temperatures than when oriented in all other directions. Thus the non-<100> orientations are normally referred to as the "soft" orientations while the <100> is referred to as "hard". The reported yield or flow stresses of polycrystalline NiAl alloys are much more varied probably due to the differences in grain size, processing history, interstitial content, texture and composition. The typical reported yield stress of polycrystalline NiAl has a strong temperature dependence below room temperature and above 800K, and is

PAGE 15

8 relatively independent of temperature between 300 and 800K (9, 45), as shown in Figure 2.5. However, the large disagreement among the data from different investigators makes it difficult to define the details (46). The data for soft NiAl single crystals are less varied while their yield stress exhibits a very strong temperature dependence below 400K, and then remains relatively constant or slowly decreases with increasing temperature (9, 11, 24, 27) (Figure 2.6). Deviations from stoichiometry increase the yield stress of soft NiAl, while the temperature dependence appears not to be disturbed as shown by the data from a [123] oriented Ni 52 Al 48 single crystal in Figure 2.6 (27). The yield stress of <100> NiAl single crystals is considerably higher than that of soft crystals and display a distinctly different temperature dependence, especially below 600K (3, 9) (Figure 2.6). Specifically, there is a strong temperature dependence below room temperature and an almost constant stress region from room temperature to about 600K. Above 600K, the flow stress drops abruptly with temperature and approaches the same level as both "soft" oriented single crystals and polycrystalline NiAl above 800K. The temperature range where this abrupt decrease in yield strength takes place is strongly dependent on both composition and strain rate. For example, Kim (27) reported that, for an off-stoichiometric Ni^Al^ crystal, this temperature range was about 200K higher than that of stoichiometric NiAl single crystals. A high strain rate sensitivity is reported for "hard" orientation NiAl single crystals tested at 473K and 1 144K (8, 9, 1 1, 28). At temperatures higher than 1 144K, the deformation is proposed to be dominated by the climb of <001> dislocations (47). 2.3 Slip Behavior The slip behavior in NiAl is an area where a lot of controversies and unanswered questions remain. This section will start with a brief summary of the theoretical efforts on the

PAGE 16

9 prediction of slip systems in NiAl followed by a summary of the experimental investigations of dislocations in both "hard" and "soft" orientation NiAl single crystals. 2.3.1 Theoretical Calculations As shown in Figure 2.1, the three shortest unit slip vectors available in the B2 structure are <001>, <1 10> and <1 1 1>. These vectors represent three perfect dislocations which will retain the ordered structure while gliding. Theoretical predictions of slip systems in this class of materials requires consideration of issues such as the ordering energy, dissociation of the <1 1 1> dislocations, dislocation line energy and mobility. In B2 materials, it is well known that deformation normally occurs by either <1 1 1> slip or <001> slip at low temperatures (48, 49). Based on the consideration that the decomposition of <1 1 1> dislocations can only occur if the separation of the <1 1 1> superpartials is small so the dislocations are compact, Rachinger and Cottrell deduced a critical value of 0.06 eV for the ordering energy (50). Below this value, the separation of < 1 1 1 > superpartials is too large so that slip can only occur via the glide of <1 1 1> dislocations. This prediction agrees well with the experimental observations that slip in high ordering energy B2 compounds such as NiAl is typically via <001> dislocations, whereas in B2 compounds with lower ordering energies (e.g., FeAl), slip is normally by <1 1 1> dislocations. However, there are some exceptions, such as AuZn and AuCd which have lower ordering energy than FeAl but still deform by <001> slip (50, 51); this may be related to other factors such as dislocation mobility. The elastic energy per unit length of a dislocation in anisotropic cubic crystals can be expressed as (52) E = • ln« An R

PAGE 17

10 while K is the appropriate anisotropic energy factor. Based on this equation, the dislocation line energy of all the three unit dislocations in NiAl have been calculated by Potter (51), Lloyd and Loretto (53), Ball and Smallman (2) and Miracle (54). Their results, as shown in Figure 2.7 and Table 2.1, were later verified in a more detailed calculation by Glatzel et al. (55). The relative mobility of various dislocations is also estimated using an equation proposed by Eshelby (56): S = 47r-exp(— — ) b b The results of these calculations are shown in Table 2. 1 . The experimentally observed slip systems in non-<100> oriented NiAl single crystals are either <001>{ 110} or <001>{010}, which is in agreement with the theoretical calculations. However, in <100> NiAl crystals where there is no resolved shear stress for <001> slip, <111> dislocations are reportedly observed at temperatures below 600K and <001> and <110> dislocations above 600K. These observations contradict the theoretical predictions, which indicate that <1 10> dislocations have lower line energies and higher mobilities than <1 1 1> dislocations. Further efforts to estimate the dislocation mobility in NiAl have been carried out by calculating the dislocation core structures (57, 58). The core structure of <001> screw dislocations obtained with central force potentials exhibits a planar extension on {110} planes (57) which predicts {110} being the favored slip plane. However, more recent work has shown that <001>, <01 1> and <1 1 1> dislocations all have non-planar cores and thus should be difficult to slip (58). Parthasarathy et al. suggested that the low dislocation mobility in NiAl could be overcome by kink movement as calculations of dislocation mobility suggest that a kinked <001> dislocation in NiAl should have much higher mobility (59). Experimentally, both the dissociation and decomposition of the <110> dislocations have been observed (60) while the cores of <1 1 1> and <100> dislocations in

PAGE 18

1 1 stoichiometric NiAl are reported to be relatively compact (61, 62). Mills and Miracle (63) reported that <110> edge dislocations exhibit a dissociated core consisting of two 1/2<111> superpartial dislocations, and some of the <110> dislocations transformed into <100> dislocations during high resolution TEM observations. The dissociation or decomposition of <110> dislocations may be the cause for the strong temperature dependence of the yield stress of <001> oriented NiAl single crystals at intermediate temperatures. Table 2. 1 Calculated dislocation line energy and mobility in stoichiometric NiAl by Ball and Smallman (2) (Data in bracket by Potter (51)). * — ln(— ) has been taken as being equal An r 0 to unity. Burgers vector Slip plane Screw/Edge E*xl0 4 S(mobility) (ergs/cm) [111] (-110) S 13.93(12.45) 0.982(0.735) E 29.3(26.90) 0.42(0.460) [110] (-110) S 10.3(10.30) 0.318(0.331) E 15.25(14.78) 0.06(0.138) [010] [-101] S 9.29(9.29) 0.46(0.480) E 8.83(8.29) 0.513(0.545) [100] [010] S 9.29(9.29) 0.67(0.653) E 7.69(7.39) 0.71(0.718) 2.3.2 Observed Slip Systems and Dislocation Behavior When oriented along soft directions, the observed slip systems in NiAl single crystals at room temperature and above are reported to be either <001>{ 1 10} or <001>{010} depending on the specimen orientation, and exclusively <001>{ 110} at 77K (2, 7, 8, 22, 64). Cube slip on {hkO} planes was also reported (25), however, which could be due to the composite cross slip of <001> dislocations between {010} and {110} slip planes as suggested by Ball and Smallman (2). It was also reported that extensive crossslip occurs when deforming soft NiAl crystals at room temperature and above (26, 27), and

PAGE 19

12 dislocation tangling, dipole and loop formation are present in the deformed specimens (24, 64). For specimens oriented along <001>, there is no resolved shear stress on <100> dislocations so that cube slip is theoretically no longer available. At temperatures at or below room temperature, <1 1 1> slip systems were observed by Sun et al. using surface slip trace analysis and TEM (29) and by Loretto and Wasilewski (22), and Bowman et al. (3) using TEM analysis. <001> dislocations were also frequently observed due to the kinking of compression specimens (65). <1 1 1> slip was not observed at high temperatures in stoichiometric NiAl crystals along the hard orientation and, instead, a high density of <100> dislocations were always observed. It is, however, generally believed that <1 10>{ llO} slip systems are active at temperatures above 600K in <001> oriented NiAl single crystals despite their infrequent observations and it has been suggested that it is the decomposition of these dislocations that produces the rather high density of cube dislocations after deformation. 2.3.3 Effect of Heat Treatment, Alloying Elements and Impurities on Mechanical Behavior The mechanical behavior of NiAl is well known for its sensitive dependence on impurities, alloying elements and thermal history. The effect of thermal history is related mostly to the formation of thermal vacancies introduced during the heating and cooling and therefore, the above-mentioned effects can be roughly divided into three categories: thermal vacancies, interstitial impurities, and alloying additions. In order to improve the low temperature ductility, efforts of alloying NiAl with ternary elements have been performed in an attempt of either creating additional active slip systems (13) or gettering of interstitial impurities (16). However, no improvement in the room temperature ductility of NiAl polycrystalline alloys via alloying was reported (66) and neither was activation of additional slip systems observed via this approach (67). The only

PAGE 20

13 report of enhanced ductility due to alloying was that by Darolia et al. (16) who reported that microalloying with Fe, Ga and Mo could increase the room temperature ductility of single crystal NiAl alloys, although the mechanism(s) for this improvement remain unexplained to date. Lower purity NiAl single crystals typically exhibit higher strength and lower ductility at room temperature. It has been proposed that interstitial impurities affect the mechanical behavior of NiAl by pinning dislocations, i.e., by reducing the density of mobile dislocations. This interaction between interstitial impurities and dislocations has been confirmed recently by the observation of strain aging effects in this material (18, 19, 20). Fast cooling from high temperature has been shown to result in supersaturation of vacancies in NiAl (68), and, upon annealing at a lower temperature, these quench-in thermal vacancies either form dislocation loops and helices (69, 70, 71, 72), or voids (73, 74, 75, 76). While higher hardness (77) and lower ductility results from the fast cooling of NiAl from high temperatures, there has been, no investigation of the slip behavior of fast cooled NiAl single crystals and consequently, it is still not clear if the supersaturation of thermal vacancies causes any change in the slip behavior. Considering that the properties of NiAl are also sensitive to the deviation from the stoichiometric composition, as shown by the effect of the stoichiometry on yield stress (Figure 2.8), and that such a deviation is by the creation of point defects since excess Al atoms will create vacancies on Ni sites and excess Ni atoms will create anti-site Ni on Al sites, the understanding of the effect of heat treatment on the slip behavior will also help understanding the effects of deviation from the stoichiometry. 2.4 Ductility and BDTT The reported temperature dependence of the ductility of NiAl alloys can be roughly divided into three stages. At low temperatures, below -500K for soft orientations or

PAGE 21

14 -600K for polycrystals and hard orientation single crystals, the ductility is usually quite limited as indicated by both low tensile elongation and low fracture toughness. The end of this low temperature stage is normally referred to as the brittle-to-ductile transition (BDT) temperature or BDTT, which is usually correlated with a dramatic increase in either tensile ductility or fracture toughness. At temperatures immediately above the BDTT, NiAl single crystals tend to exhibit high tensile ductilities (up to 300%) in the case of soft crystals and up to 80% in the case of hard crystals. At even higher temperatures, the elongation may drop again due to plastic instabilities (necking) in the specimen. 2.4. 1 Ductility of Single and Polycrystalline NiAl The ductility of <001> oriented NiAl is negligible except at high temperatures (>600K). For example, it has been reported that <001> oriented NiAl fails elastically in tension when tested below 589K (8, 28) although plastic deformation in compression was reported at temperatures as low as 77K (3, 27, 29). Similar to soft oriented NiAl, there is a dramatic increase in the tensile elongation of <001> crystals above certain temperatures (600K) and high elongations can be achieved at these higher temperatures (6, 28, 78) (Figure 2.9). This transition is strongly composition dependent, as shown in Figure 2.10. The actual transition temperature is near 600K with the scatter in the various data due to the variations in alloy purity, strain rate and composition (alloying additions as well as deviations from stoichiometry). Polycrystalline NiAl has virtually no tensile ductility (<2% typically) below about 600K. It has been suggested that the low ductility of polycrystalline NiAl is due to the insufficient number of independent slip systems provided by cube slip. Unlike single crystals, polycrystalline NiAl needs at least five independent slip systems to accommodate the grain boundary shape change according to the Von Mises criterion (79) although this requirement has been questioned by some authors (80, 81, 82).

PAGE 22

15 NiAl single crystals oriented along soft orientations usually exhibit limited tensile ductility at room temperature. For example, tensile elongations of about 1% were reported earlier for non-<001> oriented NiAl single crystals (85). However, more recent work resulted in 5-7% RT tensile elongation by adjusting the thermal treatment conditions (83, 84) or by doping with less than 0.5at% of Fe, Mo and Ga (16). Despite these improvements, the mechanism causing the low RT ductility of non-<001> oriented NiAl single crystals is still unknown since, unlike polycrystalline materials, single crystals do not need five independent slip systems to accommodate the shape change. Various mechanisms such as gettering or trapping of interstitial elements by solute atoms, dislocation core structure interactions, slip system modification, electronic effects and stoichiometry effects have been proposed to explain the increase in RT ductility via microalloying (16). Yet, none of them have been proven. In fact, the understanding of the mechanisms governing the low and intermediate temperature deformation of stoichiometric NiAl is so poor that the theoretical basis of such a microalloying approach is not clear and should be questioned. As temperature increases, it is reported that soft-oriented NiAl undergoes an abrupt brittle-to-ductile transition at intermediate temperatures and immediately above the transition temperature the tensile elongation suddenly increases to more than 100% over a short temperature range (Figure 2.9). This transition happens between 475-525K depending on the specific orientation (11, 85). The abnormally large elongation has been attributed to a balance between strain hardening and recovery processes (85). As temperature increases further, necking occurs and the plastic elongation decreases to about 20-50%. 2.4.2 Brittle-to-ductile transitions (BDT) in NiAl Unlike the BDTT in traditional metals, the BDTT of NiAl alloys does not have a consistent definition; this makes it a rather confusing terminology and often without a clear

PAGE 23

16 physical significance. In the case of single crystal NiAl, most published data about the BDTT are obtained from tensile elongation data; this may not be an appropriate representation of the BDTT. Therefore, it should be advised that the BDTT of single crystalline NiAl may actually represent a sudden increase in tensile elongation rather than the traditionally defined onset of ductile fracture. For stoichiometric NiAl, the BDTT of soft-oriented NiAl single crystals is reported to be between 475 and 525K depending on orientation (11, 25), Lhat of <001> oriented NiAl single crystals between 589K and 644K (28) and finally, the BDTT of polycrystalline NiAl is between 550K and 600K. An interesting phenomenon in polycrystalline NiAl is that its BDTT (550K-600K) is higher than that of soft-oriented NiAl single crystals even though the observed active slip systems in both types of materials are the same. Since maximally only three independent slip systems are available for <001> slip, it is normally expected that polycrystalline NiAl has very limited low temperature ductility consistent with the experimental results. However, except for the few studies where non-<001> dislocations were reported after high temperature deformation, the active slip systems are still those with <001> Burgers vectors at high temperatures; this indicates that the BDTT is not due to a sudden increase in the number of independent slip systems. Various possible mechanisms have been proposed to explain the BDT of NiAl. For polycrystalline NiAl, both the activation of new additional slip systems and climb have been suggested (53). However, the activation of new additional slip systems was not observed experimentally and neither is it consistent with the observed rate dependent deformation of NiAl near the BDTT. On the other hand, it was shown that the combination of glide and climb of <001> dislocations could result in five independent deformation mechanisms (86). Hence it has been suggested that the BDTT in polycrystalline NiAl is due to the onset of dislocation climb assisted by short circuit diffusion (66).

PAGE 24

17 It was also suggested by Noebe et al. (87) that similar climb processes are responsible for the BDTT in hard-oriented NiAl single crystals since the BDTT of <001> oriented NiAl single crystals shows a similar strain rate dependence at the same temperature range as the BDTT of polycrystalline NiAl. This view was supported by the observation of a large number of <001> dislocations in hard-oriented NiAl single crystals deformed at temperatures above the BDTT. However, the BDTT of the hard-oriented NiAl is also coincident with the slip transition from <111> to non- slip (88). Hence, the slip system transition can not been ruled out from being responsible for the BDTT in the hardoriented crystals. Some recent work has linked the BDTT of <001> oriented NiAl single crystals to the thermally activated mobility of <1 10> dislocations in that the activation of <1 10> slip systems was observed only at temperatures above the BDTT. Based on the observation of the decomposed core structure of <1 10> edge dislocations by HRTEM and the assumption that screw and mixed dislocations are relatively mobile while edge segments undergo decomposition which locks the movement of the dislocations, Mills et al. (63, 89) proposed two slightly different mechanisms all based on the diffusion-aided glide of decomposed <110> dislocations via cooperative climb of <001> dislocations as shown in Figure 2.11. However, these mechanisms rely on fast short-range diffusion which may not exist even though the local dislocation stress field and pipe diffusion along the dislocation line provide a much faster short-range diffusion than bulk diffusion at the transition temperature range. An even more severe problem with these mechanisms is that the climb of the two <001> dislocations has to be cooperative although only one of these <001> dislocations has a driving force for climb in <001> loading, which should make climb uncooperative. It is also unlikely that the decomposed <1 10> dislocations will constrict to their higher energy state in the presence of thermal activation. Thus, the mechanism for the BDTT of hard-

PAGE 25

oriented NiAl single crystals still remains undetermined, although it is generally believed that some sort of thermal activation process is responsible for the transition. The mechanism for the BDTT in soft-oriented NiAl single crystals is the one most controversial. Since the reported BDTT of soft-oriented NiAl single crystals is so low (about 0.25 T M ) and the strain rate dependence is relatively small (12), the mechanisms based on the climb process mentioned above can not be readily applied. Instead, the enhanced cross-slip, or unlocking of dislocations from pinning points have been proposed to be the mechanisms responsible, in addition to dislocation climb. There are results indicating either of these mechanisms (17, 90). However, direct observations using TEM or other methods to support any of these mechanisms are still lacking. In the case of cross slip, further details of the mechanisms must also be considered since cross slip can not increase the number of independent slip systems (24). For single crystals, the Von Mises criterion no longer holds and, therefore, it has been suggested that whether the stress concentrations in the materials can be promptly relaxed determines the level of ductility of NiAl single crystals (91). This suggestion is supported by recent work which indicates that the fracture of NiAl single crystals is crack-nucleation controlled at temperatures well below the BDTT and becomes more and more crack-propagation controlled with increasing temperature (92). 2.5 Formation of Dislocation Dipoles and Loops During Plastic Deformation It has been reported that dislocation dipoles, loops and debris are frequently observed in non-<001> oriented NiAl single crystals deformed at low to intermediate temperatures (24, 64). Similar dipoles, cusps, or loops have also been observed previously in BCC metals (e.g., in Fe-0.3Si (93)), FCC metals (94, 95), zinc (96), magnesium oxide (97) and Si (98). These features are often attributed to mechanisms such as the pinching-off of loops from the jogs in the screw dislocations. A simple mechanism proposed by

PAGE 26

19 Johnston and Gilman (99) was often used to explain the formation of dipoles and pinchedoff debris, as illustrated in Figure 2.12. This mechanism was put forward by realizing that the jogs in the screw dislocations can not glide in the same direction as the screw components. Therefore, when a jog height is small enough, the jog will be dragged to move with the dislocation leaving point defects behind. At intermediate jog heights, the sessile jog will cause dislocation dipoles to form as shown in Figure 2.12c. With superjogs, the jogged dislocation will become a single-ended dislocation source since the two edge segments can pass each other without being trapped. High densities of prismatic loops and dipoles can be readily produced via the Johnston and Gilman mechanism if the jogs can be formed via double cross slip. Arguing that the Johnston and Gilman model is too simple to account for the various and sometimes very complicated dislocation features observed in BCC and FCC metals, and noting that in many cases cross slip is unlikely unless local stress concentrations are present, Tetelman proposed another mechanism for the formation of dislocation dipoles and loops (100) as shown in Figure 2.13. In this mechanism, two dislocations of equal Burgers vector but opposite sign, slipping on parallel slip planes separated by a distance y, can trap each other by their elastic energy fields. The dislocations can then lower their energy by reorienting part of their length, and the screw segments can cross slip to annihilate and form a dipole or closed loop. This mechanism may account for the jogs and dipoles forming at the early stages of deformation where slip is confined to the primary slip planes. The jogs can also be formed via dislocation-vacancy and dislocation-dislocation interactions. However, the jogs produced by these interactions will be of very small height, and may only produce point defects when they are moving. Nevertheless, most jog forming mechanisms include a certain cross slip process. Therefore, an investigation of the details of both dipole and prismatic loop formation may result in a further understanding of the slip behavior such as the primary and cross slip planes or the intensity of cross slip during deformation.

PAGE 27

20 Al Ni Figure 2.1 The crystal structure of NiAl with the three possible slip directions shown. WEIGHT PERCENT NICKEL 0 10 20 30 40 50 60 70 80 90 W ' i ' ' ATOMIC PERCENT NICKEL Figure 2.2 Binary NiAl phase diagram (from Nash, Singleton, and Murray (33)).

PAGE 28

21 50 45 50 55 Ni (at%) 60 Figure 2.3 The elastic constants of NiAl as a function of temperature and composition, (after Rusovic and Warlimont (38)). B a i— i £ s Oh O 0289 Q288 0287 0-286 density -16 § S So Ni (%) Figure 2.4 Lattice parameter and density of NiAl as a function of composition at room temperature (after (12)).

PAGE 29

22 2000 1800 1600 o 1400 Ql w 1200 » 1000 1/1 800 .!? 600 400 200 0 0 1 T' " 1 1 Bowman •! al. 1692 1 1 1 Poseoe and Newey 1968b Grain size m 10 fim Groin Size 50/im Ni-50.GA1 O NI-53AI V Ni-49.6AI Grain size 30 pm Ni-4B.9AI A Ni-50.6A1 A Ni-43AI O Ni-4.1.6A1 s 1 -y -p — -t^' 2 ' 1 1 1 200 400 600 800 1000 Temperature (K) 1200 1400 Figure 2.5 Temperature dependence of yield stress of polycrystalline NiAl (after (12)). EQ O _ w BO — 200 400 600 800 Temperature (K) 1000 1200 1400 Figure 2.6 Temperature dependence of yield stress of NiAl single crystals for different orientations (Data compilation from (12)).

PAGE 30

23 >< E? S LU C c 3.5 3 2.5 2 1.5 1 0.5 HOP] X omi o |100](011) a [Oljj(OII) • [111 ](01 1 ) X X [Oil] ami 45 90 135 Theta (degrees) [100] 180 Figure 2.7 Line energy of [100], [Oil], and [111] dislocations as a function of dislocation character on the (Oil) plane at 933K. The horizontal axis is the angle between the dislocation line vector and the [100] dislocation. The line energy is proportional to C 44 (a 2 /47i)ln(r/r 0 ).(after Miracle (54)). 46 50 54 58 Ni (at%) 62 Figure 2.8 Yield stress of NiAl as function of composition (Data compilation from (12)).

PAGE 31

24 200r 1 1 1 1 1 r Temperature (K) Figure 2.9 Tensile elongation vs. temperature along different orientations ((6)). 100 [ : 1 200 400 600 800 1000 1200 Temperature (K) Figure 2. 10 Tensile ductility of <001> oriented Ni-50A1 and Ni-60A1 single crystals as a function of temperature (after (12)).

PAGE 32

25 Extra (0 2 2) half-plane z • [10 0] [0 1 1] "Diffusion Path" a[0 1 0] (a) Overall Direction of Motion [0 1 II Extra (0 2 2) half-plane Vacancy Flow (011) Glide Plane , t 011 (b) Figure 2.1 1 Mechanisms by Mills et al. (63, 89, 101) for the glide of decomposed <1 10> dislocations: (a) Cooperative motion of a decomposed [011] dislocation via a short-range diffusive process between two <001> dislocations. Diffusion of atoms from the tip of the lower extra (022) plane to the upper one accomplishes the glide motion of the overall [OlT] dislocations, (b) Forward propagation of a decomposed edge segment by the lateral motion of macro-kinks (MK). Glide motion of the MK requires constriction of the decomposed segments, which can occur by a conservative climb process.

PAGE 33

26 (c) Figure 2. 12 A model explaining the behavior of jogged screw dislocations by Gilman and Johnston: (a) small jog is dragged along creating point defects as it moves; (b) in the case of very large jog, segment NY and MX move independently; (c) intermediate jogs (MN) pinning dislocations and the segments NP and MO interact and can not pass each other in a process forming edge dipoles. (Figure from Low and Turkalo (93)) Figure 2.13 The mechanism for jog formation proposed by Tetelman: (a) two non-parallel dislocations MM' and NN' of opposite sign glide on parallel slip planes separated by a distance of y; (b) dislocations lower their energy by reorienting part of their lengths in the glide plane; (c) PM' cross slips down and leave a dislocation dipole MPP'R'RN' after annihilating the cross slip segment, (after Hull (102))

PAGE 34

CHAPTER 3 EXPERIMENTAL PROCEDURE 3.1 Sample Preparation 3.1.1 Single Crystal Growth Stoichiometric (Ni 50 Al 50 ) NiAl and a 0.3at%Si-doped NiAl were chosen for the purpose of this study. High purity Ni (99.99%), Al (99.999%) and Si (99.9999%) were used as starting materials. The materials were first cast into rods using non-consumable arc melting under purified argon. These arc-melted rods were then used to produce single crystal rods about 27 mm in diameter and 80 mm in length in a CENTORR Model M60 furnace under purified Ar atmosphere using a modified Bridgman growth method. High purity A1 2 0 3 crucibles were used in order to minimize contamination. After growth, the single crystals were homogenized at 1573K for 3 hours and slowly cooled to room temperature. All the specimens studied in this research are as-homogenized unless indicated otherwise. Table 3.1 Chemical composition of single crystals used in this study. Crystal Orientation Ni Al Si C S O N ID at. % at. % ppm ppm ppm ppm ppm 62 [001] 49.6 50.3 460 19 34 [557],[233] [011] 50.4 49.5 82 <5 106 <15 109 [557] 50.3 49.6 24 120 <15 64 [011] 50.6 49.3 240 180 81 <15 40 [123] 50.0 49.9 100 110 13 100 9 NiAl-Si [001] 50.28 49.44 0.25 at. % 151 <7 89 15 27

PAGE 35

2
PAGE 36

29 solution and the final gage cross section was approximately 1.3mm x 2.1mm after electropolishing to remove the recast layer and any surface roughness. The orientations (i.e., tensile axes) of specimens investigated in this study include [557], [223], [T23], [Oil] and [001], as shown in Figure 3.1b. In most cases, the Side (narrow) and Face (wide) (Figure 3.1a) were also carefully chosen to be parallel to certain crystal planes (see Table 3.2). Table 3.2 The crystal orientations of the specimens. F(l 10) and S(l 10) refer to the cases when the Face and Side planes are parallel to (1 10), respectively. Specimens Orientation Side Plane Face Plane [557] NiAl S(110) [557] (110) (7 7 10) [557] NiAl F(110) [557] (7 7 10) (110) [T23] NiAl [123] n/a n/a [233] NiAl [233] n/a n/a [011] NiAl F(110) [011] (100) (0T1) [011] NiAl S( 110) [011] (Oil) (100) [001] NiAl [001] (010) (100) [001] NiAl-Si [001] (110) (110) 3.1.5 Preparation of Transmission Electron Microscopy Specimens Foils for TEM analysis were cut with a low speed SiC abrasive wheel from the specimen gage sections either parallel to the Side or to the Face or parallel to a certain crystallographic plane in the specimen. In order to best preserve the dislocations in the specimens, TEM foils were carefully hand ground on fine grinding papers to a finish

PAGE 37

30 thickness of -150 u.m. Final thinning of the foils was done by jet-polishing using a Struers Tenupol-2 device in a solution of 70% ethanol, 14% distilled water, 10% butylcellusolve and 6% perchloric acid, at 0°C and 25V. Other precautions such as cutting TEM foils from different regions in the gage of the tensile specimens and using a low jet speed during electropolishing were employed in order to avoid specimen damage. Tensile testing was carried out on a screw-driven Instron Model 1 125 machine with a holder designed to accept the geometry of the specimens in this study. The specimens were heated using a clamshell furnace with an accuracy ±5K when tested between RT and 1273K in air. In the case of testing at 77K or 180K, the specimen and the specimen holder were either immersed in liquid nitrogen or liquid nitrogen-alcohol mixtures, and test temperatures between 77-180K were achieved by keeping specimens in the cold N 2 vapor. The temperatures were measured using a chromel-alumel thermocouple attached directly to the specimen. When testing at temperatures other than room temperature, a certain holding time was given after reaching the test temperature to ensure thermal equilibrium during testing. After deformation at elevated temperatures, the furnace was always promptly removed to ensure fast cooling of the specimens and retention of the dislocation substructures. The shear stress (t) and shear strain (y) were calculated using the following equations (103): 3.2 Mechanical Testing T = A P o sin A, 0

PAGE 38

31 7 = -sin A» cos A 0 cos \ OJ where P is the load, \ the initial cross sectional area, L, and L 0 the current and initial gage lengths respectively, 7^ the initial angle between the tensile axis (TA) and Burgers vector, and x 0 the angle between TA and slip plane. Due to the semi-brittle nature of NiAl crystals, normally at least two tests were performed for each data point at elevated temperatures. While at lower temperatures where larger scattering is expected, more than two tests were often conducted. The dislocations in the deformed specimens were investigated using a JEOL200CX TEM and a Philips EM420 TEM both equipped with double-tilt stages allowing ±60° (main) and ±30° (cup) tilts and operating at accelerating voltages of 200 kV and 120k, respectively. Conventional bright field (BF) and weak beam dark field (WBDF) images were used to determine the nature (Burgers vector and line direction) of the various dislocations. It was found that the dislocations obey the g«b criterion despite the elastic anisotropy of NiAl (53). However, whenever there was any doubt about the residual contrast at g«b=0, additional diffraction conditions were used to verify the analysis. The dislocation line directions were identified via trace analysis using standard techniques. Similarly, images from at least three zones were used to verify the results. A JEOL-35CF SEM was used to analyze the fracture surfaces and slip traces on the gage surfaces. Again, the latter were correlated with the dislocation analyses by performing surface trace analysis. 3.3 TEM Dislocation Analysis and SEM analysis

PAGE 39

32 Figure 3.1 (a) The geometry and the size (in mm) of a typical tensile specimen after EDM cutting and before electropolishing. (b) Stereographic projection with the orientations investigated in this study shown.

PAGE 40

CHAPTER 4 RESULTS AND DISCUSSION In this chapter the results of the mechanical properties of both soft and hard orientation high purity NiAl single crystals are presented. In order to understand the mechanisms of the mechanical behavior, the slip behavior was studied using TEM and SEM. The effects of thermal treatment, temperature, strain rate and orientation on the deformation is also presented. The last section consists of a general discussion of the cross slip behavior, dislocation mobility, tensile ductility and other deformation behavior of NiAl single crystals based on the understanding obtained through this work and previous research. 4. 1 Mechanical Properties 4.1.1 Soft Orientation at Room Temperature The critical resolved shear stress (CRSS), work hardening rate, and tensile elongation of a series of "soft" single crystal specimens tested at room temperature are listed in Table 4.1. The results indicate: (1) that up to 34% tensile elongation can be achieved at RT; (2) that tensile elongation is dependent on orientation and specimen geometry; (3) that there are no significant differences in the CRSS between the <001>{ 1 10} and <001>{010} slip systems for samples cut from the same crystal (the differences between the average CRSS value for each system is no more than 5% which is within the scatter); (4) the work hardening rate (WHR) is very low for orientations other 33

PAGE 41

34 than <1 10>; and (5) the fracture stresses (a f ) of all the specimens are close to each other except for that of the [123] oriented specimens which were cut from a different crystal. Furthermore, the tensile elongation also appears to be dependent on geometry, i.e., [557] specimens with the F(110) geometry (see Chapter 3) tend to exhibit higher tensile elongations than those with the S(110) geometry. This difference in tensile elongation could be caused by the difference in the specimen head constraints due to the asymmetric rectangular specimen shape (Figure 4.1), as described elsewhere (104). However, it is also possible that this difference is related to the higher CRSS of the S(110) specimens as indicated in the table. It should be noted that although the specimens of both geometries were cut from the same crystal, the F(110) specimens were actually cut from the bottom half of the crystal rod whereas the S(110) samples were cut from the top half. It was subsequently determined via chemical analysis that the compositions of these crystals were not uniform throughout the crystal, i.e., the top is normally Ni-rich due to the evaporation of Al during single crystal growth; this resulted in CRSS values about 15% higher in the [557]S(1 10) specimens. Since in lower purity NiAl crystals the yield stress is normally higher and elongation normally lower, the higher CRSS in S(110) specimens should adversely affect the tensile elongation. Typical shear stress-shear strain relationships are shown in Figure 4.2a. Once again, it can be seen that the WHR of [557]-oriented single crystals is quite low throughout the deformation. The WHR of [T23]-oriented specimens is very close to that of [557] oriented specimens suggesting similar slip behavior for both [557] and [T23] oriented specimens, while that of the [Oil] oriented crystals is considerably higher, consistent with their double slip behavior. The [T23] crystal also exhibits a lower CRSS for [001](Tl0) slip than that of [557] specimens probably due to its more stoichiometric composition (Table 3.1). The RT shear stress and shear strain relationships of these specimens suggest

PAGE 42

35 that the Stage I deformation for [557] and [T23] oriented crystals lasts until fracture, while [Oil] oriented specimens undergo stage II deformation from the onset of plastic deformation. The effect of purity on the mechanical behavior is clearly shown (Figure 4.2b) by a comparison between the results from the [557] crystals in this study with those of a previous study (64) from lower purity crystals along a similar orientation ([111]). It is obvious that the higher purity crystals used in this study exhibit higher RT ductility, lower yield stress and lower WHR. Table 4. 1 Tensile data for soft-oriented NiAl single crystals along different orientations. S=S(1 10), F=F(1 10), B denotes to the bottom half and T the top half of a crystal rod, and {010} slip while in all other cases [001](1 10) slip is assumed. A 10" 4 s" 1 strain rate was used in all the tests. Crystal ID Orient. a o. 2 MPa ^0.2 MPa WHR (6=dx/dy) MPa MPa El. % 34B [557]F .5 1 14 57 62 266 34 34B [557]F .5 115 57.5 218 18 34B [557]F .5 109 54.5 190 16 34B [233] .485 1 17 56.5 186 5 34B [233] .485 119 57.5 202 4 34B [233] .485 127 61.5 169 3 34T [557]S .5 137 68.5 102 239 15 34T [557]S .5 134 67 221 13 34T [557]S .5 127 63.5 196 10 34T [011] .5 147 73.5 205 3 34T [011] .5 133 66.5 180 3 34T [011] .5 139 69.5 192 2 64 [011JF .5 111 55.5 468 242 7 64 [011]F .5 117 58.5 563 237 6 64 [011]F .5 102 51 230 5 64 [01 1]S .5 111 55.5 448 235 6 64 [01 1]S .5 102 51 476 212 5 40 [123] .454 116 52.5 110 306 19 40 [123] .454 110 50 114 304 15 40 [123] .454 124 56.5 109 318 14

PAGE 43

36 4. 1 . 1 Soft Orientation at Non-ambient Temperatures Figure 4.3 shows the temperature dependence of the CRSS (t 02 ) of different NiAl single crystals oriented along various soft orientations. It can be seen that the CRSS of [557] oriented NiAl single crystals exhibits a strong temperature dependence below RT and decreases only gradually with increasing temperature above RT. The low temperature (below RT) CRSS of NiAl single crystals along other soft orientations was not measured. However, it is believed that the same trend is true for all the soft-oriented NiAl single crystals based on the similarities at higher temperatures. This kind of strong temperature dependence normally indicates that thermal activation is required in the movement of dislocations, such as in the case of BCC metals. It has been shown in several studies that the Arrhenius representation x 02 = x 0 exp(-Q/RT) can be successfully used to evaluate possible thermal activation processes in NiAl (46, 87). Therefore, the results in Figure 4.3 are re-plotted in an Arrhenius plot in Figure 4.4. As is apparent in this plot, for a strain rate of 10~ 4 s~', the deformation below 500K occurs by a similar process whereas, above 500K, the deformation process becomes more complicated. As shown in Table 4.1, [557]F(110) specimens have a RT tensile elongation between 16-34% when tested using a strain rate of ~10" 4 s" 1 . However, this elongation reduced to 1.3-6% when tested at 120K and increased to more than 100% when tested above 573K (Figure 4.5 and Figure 4.6). Similar to the data reported by Lahrman et al. (11) and Takasugi et al. (6), elongations of [557]F(110) specimens reach a peak value at intermediate temperatures (Figure 4.5). It is also noted that the magnitude and temperature of the peak elongation depend on the strain rate. As shown in Figure 4.5a, the peak temperature was shifted to higher temperatures with increasing strain rate. At 473 K and 1073K, the effect of strain rate on elongation actually reverses (Figure 4.5b). At 473K where plastic instability (i.e., necking) does not occur, or 573K where

PAGE 44

37 plastic instability has just occurred, the elongation decreases with increasing strain rate; this is also true for temperatures below 473K, while at higher temperatures such as 1073K where the specimens neck easily, higher strain rates result in higher tensile elongations. However, the situation where plastic instability occurs should be considered independently from that where no plastic instability takes place. Therefore, it can be concluded that at temperatures at or below 473K, the tensile elongation of soft-oriented NiAl single crystals gradually increases with increasing temperature or decreasing strain rate, i.e., there is no abrupt change in tensile ductility. At temperatures at or above 573 K, high elongations can be obtained and this high elongation is sensitive to the strain rate, consistent with the assumption that the balance between recovery processes and work hardening results in these abnormally high elongations. 4. 1 .3 Heat Treatment Effects It has been shown that the mechanical properties of NiAl are sensitive to the thermal or processing history. In this work the effect of thermal treatment is studied using high purity NiAl single crystals oriented along [557]. The single slip orientation was chosen to exclude the complicated situation caused by the interactions between different slip systems. The specimens were water quenched (WQ) or furnace cooled (FC) after having undergone the same homogenization treatment followed by reheating to 1273K. To exclude the possible effects due to specimen geometry, all the specimens tested are either with the F(110) or S(110) geometries (see Chapter 3) and only the mechanical behavior of the specimens of the same geometry were used for comparison. The RT mechanical properties of [557] oriented NiAl single crystals are listed in Table 4.2 where it is apparent that the cooling rate greatly affected the RT yield stress and tensile elongation. The CRSS of the WQ specimens was almost double that of the FC specimens, whereas the elongation decreased from about 10% in the case of FC specimens

PAGE 45

38 to less than 1.5% in the case of WQ specimens. It is believed that the substantial decrease in the tensile elongation and increase in the CRSS of WQ specimens are due to their higher thermal vacancy concentrations. Table 4.2 Influence of thermal treatment on the tensile properties of NiAl. Specimens are [557]F(1 10) single crystals. FC=furnace cooled; WQ=water quenched. Note that the WHR for WQ specimens is estimated with data at small strains where the deformation is still in the transition range before reaching Stage I deformation. Therefore, it may not Heat Treatment ^0 2 MPa o f (MPa) WHR (6=dT/dY) MPa Elongation % 1273K, FC 58.5 185 -100* 10 1273K, FC 56 173 7 1273K, FC 60 191 10 1273K, WQ 112.5 244 -300 1.5 1273K, WQ 121.5 273 1 1273K, WQ 178 <0.1 4. 1 .4 Hard Orientation Figure 4.7 shows the temperature dependence of the yield stress for high purity NiAl tested along the <001> "hard" orientation in tension. The yield stress is above 600 MPa at temperatures below 573 K and decreases rapidly with increasing temperature above 573K. Doping with 0.3 at% Si appears to increase the yield stress at temperatures below 773 K and to produce a more abrupt drop between 573 and 773 K. However, at temperatures above 773K, both NiAl and NiAl-Si specimens have similar low yield stresses. Below 473K, the tensile elongations of the specimens are typically less than 0.5% and, in many cases, the specimens fractured before yielding. The specimens typically have -2% tensile elongation at 473K and 523K, and undergo more than 20% elongation before

PAGE 46

39 fracture at and above 573K. Therefore, both the NiAl and NiAl-0.3Si single crystals tested in this work appear to undergo a rather sudden increase in ductility above 473K. The tensile elongation becomes relatively temperature independent above 773K for binary NiAl whereas the Si-doped crystals exhibit considerably higher elongations at temperatures above 673K. In order to reveal possible transitions in the thermal activation of the slip behavior, the yield stress shown in Figure 4.7 is re-plotted in an Arrhenius representation in Figure 4.9. It is obvious that a transition occurs, similar to that reported previously (46). However, the transition in the high purity NiAl appears to occur at lower temperatures and be more gradual in nature. In both Figure 4.7 and Figure 4.9, the Si-doped crystals appear to exhibit a higher transition temperature than that of the binary crystals even though this difference was not reflected by the temperature dependence of the tensile elongation in Figure 4.8. From the results of this study, it appears that the sudden increase in ductility coincides with a drop in yield stress, and that this drop in yield stress is more gradual than what was previously reported (3, 9). These characteristics indicate that a thermal activation process, rather than a rigid slip system transition, may be responsible for the yield drop and ductility increase in <001> oriented NiAl single crystals tested at intermediate temperatures. Assuming hard-oriented NiAl single crystals deform by <01 1>{01 1 } slip, the temperature dependence of the CRSS in both hard and soft orientations can also be plotted as in Figure 4.10; this shows that, similar to the results of previous researchers (12), the CRSS of hardoriented crystals drops to values approaching those of soft-oriented NiAl single crystals at high temperatures.

PAGE 47

40 Constrained (C) Less Constrained (LC) Figure 4. 1 Schematics of the rectangular tensile specimens before and after deformation. Note the shape change (shaded) is consistent with "plane strain" in that there is no change in the thickness (T) in the constrained (C) condition which corresponds to [557]S(1 10) geometry, or the width (W) in the less constrained (LC) specimens which corresponds to [557]F(1 10) geometry. TA is the tensile axis, b the Burgers vector and n the slip plane normal.

PAGE 48

41 Figure 4.2 (a) RT shear stress vs. shear strain for NiAl single crystals having different orientations obtained using a strain rate of 10" 4 s 1 . (b) True stress vs. true strain for [5571 crystals in this study along with the results from [111] crystals by Field et al (64)

PAGE 49

42 73 273 473 673 873 1073 1273 Temperature (K) Figure 4.3 Temperature dependence of the CRSS at a strain rate of 10" 4 s" 1 for [557] and [123] oriented crystals ([001](1 10) slip system) and for [Oil] oriented crystals ([001](010) slip system). 1000 i , 100 TO Q(I) V) o 1 o 0 20 40 60 80 100 120 140 1 0000/Temperature(K) Figure 4.4 Arrhenius representation of the temperature dependence of the CRSS for softonented crystals.

PAGE 50

43 c o re o> c o Q) 0) '5 c 200 180 160 140 120 100 80 60 40 20 0 373 — SR 10-2 — SR 10-3 SR 10-4 573 773 Temperature (K) (a) 973 1 173 180 160 140 C o 1 20 re D) 100 c o
PAGE 51

44 100 200 300 Temperature (K) 400 500 Figure 4.6 Tensile elongation of [557]F(1 10) NiAl single crystals at low to intermediate temperatures tested at a strain rate of 10" 4 s" 1 . 1200 173 373 573 773 Temperature (K) 973 1 173 Figure 4.7 Temperature dependence of tensile yield stress of <100> oriented NiAl single crystals. NiAl-Si was doped with 0.3 at% Si. It should be noted that at temperatures below 573K, the data may represent the fracture stress rather than the yield stress because the activation of <1 1 1> or <1 10> slip systems could not be verified by dislocation analysis

PAGE 52

45 80 t > 70 sS 0^ b0 c 0 50 re D) c o LU 30 oriented NiAl single crystals. 10000 CD 1000 oriented NiAl single crystals.

PAGE 53

46 600 500 ~ 400 re a. 2 in tr o 300 200 100 <011>{011) <001>{110} 0 200 400 600 800 1000 1200 1400 Temperature (K) Figure 4. 10 Comparison of the temperature dependence of the CRSS of [001] and [557] oriented NiAl single crystals. The <01 1>{01 1 } slip systems are assumed in calculating the CRSS for the [001] orientation.

PAGE 54

47 4.2 Slip Behavior of Non-<001> Oriented High Purity NiAl Single Crystals at RT 4.2.1 T5571 Crystals The [557] orientation was selected for the initial investigation of the slip behavior in high purity NiAl single crystals due to the expected single slip ([001](110)) nature corresponding to this orientation. At first, the microstructure of the as-homogenized crystals was checked and as shown in Figure 4.1 1, the initial density of dislocations in the undeformed crystals is extremely low. In order to determine the slip system and fundamental slip behavior, a [557]S(1 10) specimen was deformed to -3% strain and foils for TEM analysis were cut both parallel and perpendicular to the active (TlO) slip plane. Figure 4.12 shows the dislocation substructure in the foils cut parallel to (TlO) along with the dislocations. It can be seen that the substructure is not uniform with tangled dislocations in some regions, which, nevertheless, indicates the easy glide of dislocations in this specimen. It will be shown later in this section that this kind of tangled substructure is actually parts of three dimensional patch-like discrete dislocation bands as illustrated in Figure 4.19. It should be pointed out that the dislocation bands in this study refer to the above-mentioned patch-like dislocation substructure. The results of g.b analysis indicate that essentially all of the dislocations have the expected [001] Burgers vector and that little if any secondary slip occurred. Although most dislocations have a mixed character, the trace analysis shows that the average line direction is approximately [1 10] especially in the tangled regions. This is the edge direction for [001] dislocations on (TlO) planes. Two primary features of the dislocation substructures were characteristic in these specimens. First, most dislocations are bowed out and tend to have a mixed character about 40-50° from the edge orientation, while almost no screw or close-to-screw segments were observed except for some very short segments (Figure 4.12). Second, in some

PAGE 55

48 regions, the dislocations are heavily pinned and form edge dipoles sometimes leaving traces of pinched-off loops often elongated along the [1 10] direction (arrow points, Figure 4.12). These dislocation features are important in understanding the dislocation behavior in NiAl as will be discussed below. Figure 4.12a shows that the bowed-out dislocations tend to have sharply bent tips. The sharp bent feature of dislocations in NiAl was attributed by Loretto and Wasilewski (22) to the tendency of dislocations to avoid pure screw orientations due to the relatively higher elastic energies in screw orientations. However, it is unlikely that a long screw segment could relax into much longer 40-50° segments during unloading since there is very little if any decrease in the total energy by going through such relaxation. This can be seen by calculating the total energy change AE of a unit length of screw dislocation relaxing into segments with an angle of 6 between the line direction and the Burgers vector (Figure 4.13a) AE = E(6) • L(d) E(0) • L(0) = ^ £(0) COS0 while E(0) is the energy of a screw dislocation, E(0) the energy per unit length of the relaxed segments, which are shown in Figure 4.13. A preference of [001] dislocations to lie along <1 1 1> directions (55° or 125° from [001]) was also reported by Kim (27) and was attributed to dislocation line relaxation because these directions he close to the theoretically-calculated, lowest energy configuration (2, 105) as shown in Figure 4.13. Nevertheless, the bowed-out dislocations with the sharp tip in Figure 4.12 could also be due to either the differences in the mobility of different [001] segments or pinning by small jogs at the screw segments. As illustrated in Figure 4.14, the dislocations bow out under stress and if the screw segment has the highest mobility, a tip and 40-50° segments could form. It has been suggested by various workers that the <001> screw dislocations have higher mobility than <001> edge dislocations in NiAl, especially at low temperatures,

PAGE 56

49 based on the fact that <001> screw dislocations have never been observed. In fact, theoretical calculations of dislocation mobility using Eshelby's equation (56) and core structure using embedded atom methods (EAM) all suggest that <001> screw dislocations have higher mobility (106). The other features such as the zigzag nature of some dislocation segments and the density and elongated nature of <001> loops and the various dislocation debris provide further information regarding the dislocation slip processes. Figure 4.15 (a) and (b) exhibit higher magnification images of a zigzag dislocation segment present in Figure 4. 12. The b and d segments lie in the (TlO) slip plane approximately along <111> directions and therefore, have mixed character, while both the a and c components have line directions close to [010] and lie in the (100) plane. This means that a and c are both edge jogs which can only slip on the (100) plane. The Schmid factor for the [001](100) segments is about 0.35 at the [557] orientation and, therefore, these jogs should experience a lower shear stress. These features are illustrated schematically in Figure 4.15 (c). It is also interesting to notice that the primary slip plane segments b and d have a straight <111> line direction which is not common among those dislocations which do not have cross slip segments. However, according to Figure 4.13, the <001> dislocations on {TlO} planes have the lowest line energies in directions close to [1 1 1] and [1 IT], in good agreement with the directions of b and d. This indicates that when the dislocations are pinned by closely spaced cross slip segments, they tend to align themselves in their lowest energy directions. This tendency is not as easily observed when the dislocation lines are pinned by cross slip segments that are separated by a distance larger than a critical value, since these dislocation segments will tend to bow out under the stress field. A high density of small prismatic [001] loops are observed in the deformed NiAl as shown in Figure 4.12a. An interesting feature is that most of the loops are elongated along directions close to [01 1] which should indicate the mechanism for their formation and the

PAGE 57

50 nature of the slip process. Since the densities of grown-in dislocations or annealing loops are extremely low in the NiAl crystals used in this study, as shown in Figure 4.1 1, there is no doubt that these prismatic loops are produced by plastic deformation, and that these loops are presumably formed by some sort of drag mechanism. Indeed, TEM images from deformed specimens revealed the presence of pinched-off loops from the trailing end of edge dipole segments (Figure 4.16). Such features could be produced by the pinning of screw dislocations by sessile jogs according to the Johnston and Gilman model. It is assumed that the [001] dislocations readily cross slip upon encountering obstacles as it has been suggested that cross slip is profuse in "soft" orientation NiAl (2, 8, 27). Thus, high densities of jogs could form via double cross slip. The mechanism proposed by Tetelman could also result in the observed prismatic loops elongated along [1 10] although it requires more restrained situations and may be only responsible for the dipoles or loops at the initial stage of deformation or at the regions between slip bands where cross slip of dislocations is less common. Based on the above discussion, the formation of [001] prismatic loops in [557] oriented NiAl single crystals can be explained based on the double cross slip mechanism as illustrated in Figure 4.16(c). After the dipoles and jogs are formed via this double cross slip process, the pinching-off of the prismatic loops will proceed by the Johnston and Gilman mechanism (Figure 2.12) if the heights of the jogs are appropriate. After having considered the mechanisms of prismatic loop formation, the differences between the densities of loops in and out of slip bands may now be explained as the different scale of cross slip activities in and out of the slip bands. This result also indicates that cross slip is relatively easy in NiAl. However, in high purity NiAl, cross slip may only take place when dislocations encounter the stress fields imposed by other dislocations. This also indicates that very few other obstacles exist for dislocation movement in the high purity NiAl single crystals used in this study.

PAGE 58

51 In order to study the dislocation activities outside the tangling region and investigate the characteristics of the dislocation bands, TEM foils perpendicular to the (TlO) slip plane and parallel to (001) were also made and investigated. Although only residual contrast exists in the TEM diffraction contrast images when the specimen is viewed with the Burgers vector [001] parallel to electron beam direction, as in the case of Figure 4.17, it is still obvious that the dislocation bands lie along [1 10] directions in these specimens, which verifies the [001] (TlO) slip system. It is also noted that the residual contrast is strongest when g=Tl0 is used to form the image; this is due to the fact that the average direction of the <001> dislocations is close to the edge direction (i.e., u=[110]), and for edge dislocations to have total invisibility, both g.b=0 and g.bxu=0 must be satisfied. As shown in Figure 4.17, using g=l 10 resulted in the weakest residual contrast, which is consistent with the edge line direction of [001] dislocations on (TlO) slip plane, i.e., u=[110]. In Figure 4. 17, it can also be seen that the projection of dislocation segments which are not in band regions are usually straight and lie along [1 10]. This implies that these dislocations are strictly confined to the slip plane. However, the dislocation bands are broadening or bowing out from the [1 10] direction in many places. This suggests either intensive double cross slip or activation of new dislocation sources in the neighboring (TlO) planes of the already deformed layer. The latter is less likely because such activation of new dislocation sources requires the existence of high stress concentrations which is not in agreement with the low WHR (Table 4.1). It has been suggested that cross slip is responsible for the broadening of slip bands or dislocation bands in many other materials, such as Fe-Si single crystals (107), copper alloys (108), neutron-irradiated Cu (109), and even h.c.p. crystals where cross slip is less frequently observed due to the splitting of dislocations into partials (1 10). In the case of [557] NiAl, the slip is planar due to its single slip nature. However, massive micro-scale

PAGE 59

52 cross slip occurs in this material as indicated by the work of Ebrahimi et al. (26), Takasugi et al. (25) and Ball and Smallman (2). The distribution of prismatic loops in the deformed [557] specimens also indicates that massive cross slip occurred in or near the dislocation bands. Therefore, such broadening of the dislocation bands in NiAl is likely to be caused by double cross slip of [001] dislocations. The process is essentially the same as that illustrated in Figure 4.16 for formation of the prismatic loops and jogs, and is described in further detail below. At first, the screw component of a dislocation at the edge of a dislocation band cross slips under the stress field of the high density of dislocations inside the band. After moving on the cross slip plane for a certain distance h, which is a probability parameter depending on the local stress and temperature, the dislocation cross slips back onto a neighboring primary slip plane which is relatively free of dislocations. This will create a jog of height h, as illustrated in Figure 4.16. If h ho, a Frank-Read source is generated and, during continued straining, will produce loops and dislocations gliding on this new plane until further obstacles are met. In general, the dislocations inside the slip bands are pinned by numerous jogs as well as the high density of debris and loops and, therefore, are less mobile. The continued deformation of the specimen is mainly carried out by the new sources created at the edge of the dislocation bands and these bands become broadened as the amount of deformation increases.

PAGE 60

53 Due to the formation and broadening mechanisms of the dislocation bands, their character and the overall substructure may reveal the role of cross slip in this material which may be important to understanding the mechanical behavior of NiAl. In the following sections, the effects of thermal treatments and deformation temperature on the dislocation substructure will be investigated in light of this understanding. The dislocation substructure in more highly deformed [557] crystals (-10%) are shown in Figure 4.18. It can be seen that essentially all the dislocations are still the [001] type consistent with the stage I deformation exhibited by these crystals prior to fracture. The substructure consists of numerous discrete bands which tend to be thicker than those after 3% deformation, although the distance between the bands remains about same. Figure 4. 1 8b shows that the trace of these dislocation bands lies approximately parallel to the Face of tensile specimens, which indicates these bands are still parallel to the (TlO) slip plane. This distribution is consistent with the development of the dislocation bands observed in the specimen deformed 3% and it indicates that the slip in these specimens is localized in the form of discontinuous dislocation bands lying parallel to the slip plane as illustrated in Figure 4.19. This can be explained by the broadening of the dislocation bands discussed above. These bands appear to develop from small scale dislocation tangling which initially could be caused either by impurity obstacles in the material or by interaction between opposite dislocations slipping on neighboring slip on planes. 4.2.2 fl231 crystals The dislocation substructure in a [123]-oriented specimen after 3% deformation is shown in Figure 4.20. It can be seen that dislocation bands similar to those developed in [557] specimens are also formed in [T23] specimens. Since it was reported that the surface slip trace indicated {hkO} slip plane for [123] oriented NiAl single crystals (25), special attention was made to determine the slip plane of the dislocations. The results of g.b and

PAGE 61

54 trace analyses verified that the majority of the dislocations have b=[001], and the dislocation bands lie along u=[l 10]. Based on the knowledge gained from the dislocation substructure in [557] oriented specimens, we can expect a [001](T 10) slip system operating in this specimen. Other characteristic features in the [557] specimen are also present such as the (TlO) plane (primary slip plane) view which shows debris and loops elongated or stretched along [1 10] (Figure 4.20a). A BF TEM image taken with the beam direction near [001] shows the slip plane edge-on (Figure 4.20c), as can be seen, some dislocations are bowed out of their slip plane suggesting cross slip activity. However, the dominating dislocation line direction and the edge-on trace of the (TlO) slip plane unambiguously indicate that the primary slip plane is still (TlO), rather than the maximum resolved shear stress (MRSS) plane which should have an edge-on trace ~ 20° away from that of (TlO) in the B~[001] images. Nevertheless, a surface slip trace of the MRSS plane could be produced if massive scale composite slip on the primary (TlO) and secondary (010) slip plane occurs; this is more likely in the near-surface regions due to the complex stress state there. 4.2.3 [011] Crystals The dislocation substructure in [01 1] -oriented NiAl resembles that in the specimens oriented for single slip in that it is inhomogeneous in nature. Dislocations in [01 1] -oriented specimens exhibit a patch-like distribution when TEM foils are cut parallel to one of its slip planes, as shown in Figure 4.21 . The slip vectors have been verified to be [001] and [010]. When viewing with the beam direction parallel to (010), as shown in Figure 4.21a and b, only [001] dislocations are visible. However, from the residual contrast, the projections of [010] dislocations can also be seen and appear to be tangled with the [001] dislocations, which suggests frequent interaction between the two slip systems. The [001] dislocation bands tend to lie along [100], as does the high density of prismatic loops which are located

PAGE 62

55 Figure 4. 1 1 TEM BF image showing the typical microstructure of undeformed crystals. Figure 4 12 Dislocations in a [557]S(1 10) specimens after 3% tensile strain at a strain rate of 10 s : (a) g=002 near B=[l 10], (b) g=01 1 near B=[100], and (c_) g =020 near B=[100] showing invisibility condition. The foil was cut parallel to (110) slip plane.

PAGE 63

56 Figure 4. 12 continued

PAGE 64

57 e L(9)=1/cos(6) e L(0)=1 (a) 1/E(9) 1.2 Figure 4.13 (a) Illustration of change in dislocation length after the relaxation (dashed line) and (b) Inverse of elastic energy E(0) of <001> dislocations on { 1 10} planes in NiAl (Data compilation from (27)). 0 is the angle between the dislocation line direction and its Burgers vector. 6=90° corresponds to the edge orientation while 9=0° or 180° corresponding to screw orientations.

PAGE 65

58 Figure 4. 14 Schematic illustration of the forming of the bowed-out 40-50° dislocations and sharp tip. The dislocation will bow out in the edge direction due to higher mobility of the screw segments. The sharp tip could be due to relaxation of the short screw segment. Figure 4. 15 Higher magnification images of the zigzag dislocations in (a) Figure 4 12a g=002 near B=[l 10], and (b) Figure 4.12b, g=01 1 near B=[100]. (c) Illustration of the zigzag segments.

PAGE 66

59 Figure 4. 15 continued (a) Figure 4.16 TEM images showing pinched-off of prismatic loops (a and b) and (c) a schematic illustration of the formation mechanism.

PAGE 67

Figure 4. 16 continued

PAGE 68

61 Figure 4.17 Dislocations in a [557] S(110) specimen. 3% strain at a strain rate of 10" 4 s" 1 Foil face parallel to (001): (a) g=l 10, strong residual when g.bxu * 0, (b) g=020, g.bxu * 0, (c) g=l 10, weakest residual contrast when g.bxu = 0 and (d) g=200,'g.bxu * 0. Images taken using a similar deviation parameter s.

PAGE 69

*HHWmHBlHIHMMWiHhWM>»>»l>OmiW»n jjj Figure 4. 17 -continued

PAGE 70

. MMMMMMHMttMMBtMtOlMtaM 63 (b) Figure 4.18 Dislocation substructures in a [557]S(1 10) specimen after 10.3% deformation. Foil normal about 15° from (1 10) slip plane normal. BFTEM images showing: (a) dislocation bands, (b) the dislocation tangling with B near (110), and (c) the invisibility condition.

PAGE 71

64 (c) Figure 4.18 — continued (110) Figure 4.19 Schematic illustration of the distribution of dislocation bands in [557] oriented NiAl single crystals deformed at RT. (1 10) is the slip plane. It can be seen that the slip is localized in the form of patch-like dislocation bands.

PAGE 72

65 Figure 4.20 Dislocations in [123] oriented specimen after 3% deformation: (a) (1 10) slip plane view, (b) (1 1 1) plane view and (c) (001) plane view.

PAGE 73

Figure 4.21 Dislocations in [01 1] oriented NiAl after 6. 1% deformation at RT showing: (a) dislocation bands along [100], (b) prismatic loops elongated along [100] and jogged dislocations , and (c) dislocation bands when B near [011].

PAGE 74

Figure 4.21 -continued

PAGE 75

Figure 4.22 Illustration of proposed double cross slip mechanism for forming dipoles and pinched-off loops in [01 1] oriented NiAl single crystals.

PAGE 76

69 primarily within the bands. Dislocations are often severely jogged and zigzagged, as shown in Figure 4.21b, which could be due to the pinning of either jogs or interacting dislocations from another system, or dislocation relaxation due to the higher energy of the screw dislocations. Since [100] is the edge direction of both [001](010) and [010](001) dislocations, the characteristics of the dislocation substructure are similar to those observed in [557] oriented crystals. Thus, it is clear that similar dislocation behavior exists in both single and double slip oriented NiAl crystals. Based on the similar analysis that was used to determine the slip plane of [557] and [T23] crystals, the slip systems in [Oil] crystals can be determined to be [001] (010) and [010](001). Since [001] and [010] dislocations always tangle together, as shown in Figure 4.21b, the interaction between these two systems may be the origin of dislocation tangling and bands. However, the high density of prismatic loops indicates that cross-slip is common, as it is unlikely that such jogs are produced via interaction between the two kinds of dislocations. Similar to the dislocation substructure in the single slip specimens, these dislocation features can be explained using a similar double cross slip mechanism (Figure 4.22) as the one for the single slip orientations; this predicts the formation of pinched-off loops elongated along [100] from the edge dipoles created via double cross slip. Therefore, in [011]-oriented crystals, both the [010](001) and [001](010) slip systems are operating. The interaction between [001] and [010] dislocation is responsible for the higher work hardening rate (Figure 4.1.1) and may create initial dislocation tangles which subsequently broaden via cross slip of both kinds of dislocations. 4.3 The Effect of Heat Treatment on Slip Behavior at Room Temperature Two specimens from the fractured furnace cooled (FC) and water quenched (WQ) [557] S(110) NiAl tensile specimens were chosen for dislocation substructure analysis.

PAGE 77

70 The TEM foils were cut parallel to the Face plane. The WQ specimen was deformed to about 2% plastic elongation before fracture and the FC specimen -8%. The dislocation substructure in the FC [557] specimen consisted of bands (Figure 4.23) similar to those observed in the [123] and [557] specimens described in Section 4.2, resulting from the double cross slip of dislocations on (TlO) planes. BFTEM images taken with a beam direction near [001] clearly show the images of the dislocation bands edge-on (Figure 4.23a). It can be seen that these bands are broadening in some regions although some bands exhibit a relatively straight edge. As already discussed in Section 4.2, this kind of broadening is caused by intensive double cross slip from the heavily deformed primary slip plane to the relatively dislocation-free neighboring primary slip planes. It is important to note that there is very little slip activity in the regions between slip bands (Figure 4.23a). A slip-plane view of the dislocations was also obtained and the following features are obvious (Figure 4.23c): (1) most dislocations are in the band regions and (2) a high density of debris and prismatic [001] loops elongated along [110] inside these dislocation patches (arrows). These features are similar to what was observed in the [557] oriented specimens discussed in Section 4.1. and, if there is any difference, it is that there is an even lower density of dislocations in the regions between the slip bands. Therefore, it appears that furnace cooling did not affect the slip behavior in the [557] oriented specimens as compared with the homogenized specimens studied in Section 4. 1 , except that there are even fewer obstacles for dislocation glide in FC crystals as indicated by the dislocation-free regions between the dislocation patches. The higher density of debris in the bands and the thicker bands indicate that the deformation occurred by localized slip, i.e., by the cross slip of dislocations. In the WQ specimen however, a radically different dislocation structure was observed as shown in Figure 4.24a. It is obvious that the dislocation structure is much more uniform compared with that in the FC specimen. Although there is still slip

PAGE 78

71 localization as indicated by the dislocation tangling, debris and small loops distributed throughout the regions between the bands (Figure 4.24c-f), the structure suggests more uniform slip processes than in the FC specimen. Since it has been shown that the primary mechanism for forming debris and small prismatic loops in NiAl is by double cross slip, the dislocation substructure in Figure 4.24 indicates that cross slip was occurring everywhere, i.e., outside as well as within the bands; this indicates a more interrupted dislocation movement in the WQ crystals. It can be seen that the loops and most of the dislocations are in contrast with g=10l and g=002, and out of contrast with g=l 10 and g=200. Consequently, the Burgers vector of the majority of the dislocations and loops is still [001]. However, unlike in the FC crystals, there are a small fraction of dislocations belonging to a secondary vector in the WQ crystals, as shown in Figure 4.24b. g.b analysis shows that these dislocations are either [100] or [010] dislocations. Tilting the TEM specimen to the [001] zone places the slip plane edge-on as shown in Figure 4.24d. Once again, the numerous loops spread out over the region and it again shows that slip in WQ specimens is less localized than in FC specimens and other single slip oriented specimens studied in previous sections. Besides showing some traces of primary dislocation bands in the [1 10] direction, which indicates the primary slip system is [001](Il0), the edge-on image in Figure 4.24g also shows straight traces of [001] dislocations lying perpendicular to the (TlO) slip plane (B2 in the figure) and traces of prismatic [001] loops along both the [110] (LI in the figure) and [010] (L2 in the figure) directions. Except for the LI loops, other features were not observed in the slower cooled specimens. The double cross slip mechanism of [001](Tl0) dislocations proposed for the formation of [001] prismatic loops elongated along [110] directions can not be used to explain the loops elongated along [010] in the [001] projection. Illustrated in Figure 4.25, one possible mechanism of forming these traces is the slip by a <001>{010} system which

PAGE 79

72 could leave debris or prismatic loops elongated along {100} directions via double cross slip. The straight projection trace of [001] dislocations along the trace of [1 10] planes can only be explained by slip on [1 10] planes, which is not understandable at this time as there is no resolved shear stress for the [001](1 10) slip system. Nevertheless, all these features suggest that cross slip in NiAl is common and not only confined to the dislocation band region as in the case of FC or homogenized [557] crystals. Apparently, the water quench changed the micro-scale slip process of [557] oriented NiAl single crystals. This may represent changes in the dislocation mobility and cross slip behavior in the vacancy hardened NiAl. More work will be needed to address these issues further. The difference in the amount of the plastic deformation (8% vs. 2%) that the FC and WQ specimens experienced before fracture may also account for some of the differences in the dislocation substructure. However, it is unlikely that these differences in the prismatic loops and the occurrence of secondary slip are due to the difference in the amount of deformation, since these prismatic loops and the secondary dislocations should not disappear at RT once they are produced. It is, however, possible to explain some of the changes in the mechanical behavior of the WQ NiAl single crystals with the dislocation structures revealed here. Compared with slowly cooled specimens where the glide of [001] dislocations is easy at the beginning of deformation and where no cross slip will take place in the earliest stages, it appears that the [001] dislocations in the WQ specimens have difficulty gliding from the beginning and that cross slip is induced prior to any interaction with the stress fields of other dislocations. The fact that the prismatic loops are no longer only elongated in one direction suggests that slip may not be as confined to the primary (TlO) slip plane. The lesser tendency to form dislocation bands (or localized slip) could also be attributed to the difficulty in the movement of cross-slipped dislocations as was shown in a previous section where the broadening of the bands is assisted by the cross slip of dislocations. Due to the difficulty

PAGE 80

73 Figure 4.23 The dislocations in the FC [557] oriented NiAl single crystals tested to 8% at a strain rate oflO" 4 s 1 : (a) g=200 and B near [001], (b) g=101 and B=[l IT], and (c) g=002 and B near [1 10]. Arrows point to elongated loops.

PAGE 81

74 Figure 4.23 — continued 7t •ft 4T i • •< V *'t J. f ft p=not 6:s \.im > (a) Figure 4.24 TEM BF image showing dislocations in the WQ [557] oriented NiAl single r#nw^ (a) ^n 01 ^ B near [1 1 1] ' (b) S =1 10 and B "ear [TTl], (c) g=002 and B near [1 10J, (d) g=200 and B near [001], residual contrast, (e) high magnification of (c) (f) g=002 and B near [1 10], (g) high magnification of (d).

PAGE 82

Figure 4.24 — continued

PAGE 83

76 (e) Figure 4.24 continued

PAGE 84

77 Figure 4.24 -continued

PAGE 85

78 Figure 4.25 The illustration of the possible mechanisms which can explain the dislocation features shown in Figure 4.24g: double cross slip of [001](1 10) dislocations producing [001] prismatic loops LI; double cross slip of [001](010) dislocations resulting in L2 loops;_[001] dislocations slip on (1 10) plane resulting trace B2; and Bl is the primary [001](1 10) dislocation bands.

PAGE 86

79 in the motion of cross slipped dislocations, the stress concentrations may develop to such a high value at dislocation tangles that even secondary dislocations ([010] and [100]) are generated (Figure 4.24b and d). It is also apparent that, in the WQ specimens, the stress concentration will develop more rapidly at localized strain points with much smaller strain than in the slowly cooled specimens, as also indicated by the difference between the work hardening rates of the WQ and FC specimens (Table 4.2). Hence, the room temperature tensile plasticity is greatly reduced. This behavior could be readily attributed to the dislocation-vacancy interactions caused by supersaturated thermal vacancies in the WQ specimens, consistent with the much higher CRSS for the WQ specimens (Table 4.2), although dislocation pinning by vacancy clusters is also possible as indicated by the shear loops in the slip plane (Figure 4.24f). However, the detailed mechanisms still need to be investigated. 4.4 Effects of Temperature and Strain Rate on the Slip Behavior of Soft-Oriented NiAl Dislocations in tensile specimens oriented along [557] and tested at either 120K or 77K are shown in Figure 4.26. As shown in Figure 4.26a-c, standard g.b analysis indicated that the dislocations have the same [001] Burgers vector as in the RT deformed [557] crystals. However, unlike the dislocations in the [557] specimens deformed at RT, the dislocations after sub-ambient deformation tend to be long and relatively straight, especially in the specimens tested at 77K (Figure 4.26e). The projected images of the dislocations are consistent with a true line direction u T =[110]; this is consistent with a [001](Tl0) slip system and indicates edge or near-edge character for the majority of the dislocations which suggests an even lower mobility for the <001> edge components. The dislocation substructure in the specimen deformed at 120K displays similar features such as zigzag dislocation lines, edge dipoles and prismatic loops elongated along [110] (Figure 4.26d). These features suggest that the cross slip of dislocations is also taking place at

PAGE 87

80 these lower temperatures. However, the absence of dislocation patches and the lower density of loops suggest that the cross slip activity at 120K is not as intense as at RT. This kind of dislocation substructure is consistent with the Arrhenius representation of the yield stress (Figure 4.4) which indicated no change in the slip behavior at temperatures below RT. Since the Arrhenius representation of the yield stress suggests a change in slip behavior at temperatures above 500K, and since it has been reported that there is no change in the slip systems in soft-oriented NiAl single crystals, it is of interest to know what kind of changes occurred in the dislocation substructure after deformation at these higher temperatures. Figure 4.27 shows the dislocation substructure in a [557] crystal deformed to about 18% deformation at 473K using different strain rates. It can be seen that, in general, the dislocation distribution resembles that in RT deformed [557] specimens, which indicates no substantial difference between RT and 473K slip behavior. However, it appears that the strain rate does influence the dislocation substructure. The dislocations in the specimen tested at the higher strain rate appear more stretched out and lying closer to the edge orientation which suggests either the difference between the mobility of screw and edge dislocations became larger, or the screw dislocations were dragged more severely by the sessile jogs. Since the mobility of unpinned or unjogged dislocations is unlikely to be affected by strain rate, it is more probable that the above feature is due to more severe jog dragging with increasing velocity of the screw dislocation. It can also be seen that the stress concentration at the dislocation bands in the higher strain rate specimen appears to be larger since the activation of secondary slip vectors occurs (Figure 4.27b) in contrast to the behavior in the lower strain rate specimen (Figure 4.27d). This may be due to the fact that the stress concentration at the front of the slip bands could not be readily released by cross slip or by generation of new dislocation sources via cross slip, since the movement of the

PAGE 88

81 jogged screw dislocations and the cross slip activity can not keep up with the higher imposed strain rate. As a result, the specimens tested at the higher strain rate tend to have lower tensile elongations at this temperature as shown in Section 4. 1. The scenario changes somewhat at even higher temperatures (e.g., 673 K) where the specimens can be deformed easily to very large elongations. Figure 4.28 exhibits the dislocation substructure in a [557]S(1 10) specimen deformed at 673K. Since at this temperature the tensile specimens form necks, the samples cut from the necked segments of the tensile specimen should have undergone higher local strains than the other parts of the specimen. It can be seen that there are some features in the dislocation substructure resembling the dislocation features in RT deformed specimens such as the high density of loops and debris (Figure 4.28a) and jogs pinning the dislocations (Figure 4.28c). However, the existence of dislocation bands is less obvious and the deformation is more uniform than that at RT and 473 K. Furthermore, some sort of sub-boundary has already started to form at this temperature (Figure 4.28b) suggesting that the diffusion rate has reached a level that allows climb to assist in the recovery process. The recovery of the dislocation substructure also is indicated by the relaxed dislocation lines and the larger average size of the prismatic loops, as shown in Figure 4.28c. The dislocation substructures in Figure 4.28 should represent the slip in the later stages of the deformation where necking causes inhomogeneous deformation throughout the gage of the specimen and, for those dislocation substructures formed during the earlier stage of deformation, there was sufficient time to allow recovery of the structure. To investigate the slip behavior in the earlier stage of deformation, a [T23] crystal was deformed to 3% at 723K before the TEM samples were cut from it. Figure 4.29 shows the dislocation substructures in this specimen. It is obvious that, at the earlier stage of deformation, both slip localization and formation of dislocation bands occur. However, the bands are much more broadened and much less dense than those usually observed at RT.

PAGE 89

82 The higher magnification image of the dislocation bands in Figure 4.29b clearly shows that extensive cross slip occurred and the traces of most of the cross slip segments lie along the [200] direction (arrows), which suggests a (010) cross slip plane. The above results indicate that although the slip behavior at these intermediate temperatures still remains the same (i.e., the slip is still localized and the slip system is the same), the local strain and stress concentration can be released more effectively probably because of easier cross slip, higher mobility of the jogged screw dislocations (due to faster short-range diffusion), and faster recovery of the dislocation substructures due to climb. Once again, by testing at even higher temperatures, the only active slip vectors in soft-oriented NiAl single crystals are <001> type. The slip vectors other than [001] are present presumably due to the existence of necking in the specimen gage which changes the local stress state. The highest elongation could be achieved at certain temperatures depending on the strain rate and orientation, due to the balance between strain hardening and recovery (6, 112). Figure 4.30 shows the dislocation substructures in the crystals deformed at 1073K at strain rates of 10" 4 s" 1 and 10"V. It can be seen that the deformation at elevated temperatures in soft NiAl crystals is controlled by both the glide and climb of cube dislocations. In the case of the slower strain rate, large dislocation networks and subboundaries form in the specimen (Figure 4.30a-b). These dislocation networks often consist of only [001] dislocations (Figure 4.30a) indicating the occurrence of a considerable amount of climb. In the case of the faster strain rate, the dislocation networks are comprised of two types of <001> dislocations interlaced together (Figure 4.30c); this indicates that the contribution from climb decreases at the higher strain rate as expected. The tensile elongations of the corresponding specimens can also be explained with the dislocation behavior shown in the figure. Since the amount of recovery in the specimen deformed at a strain rate of 10" 2 s 1 was lower, the specimen exhibited better resistance to

PAGE 90

83 Figure 4.26 Dislocations in [557] oriented NiAl deformed -1.3% at 120K (a-d): (a) g=002 near B=[100], (b) g=01 1 near B=[100], (c) g=020 near B=[100] showing ' dislocation out of contrast and (d) g=002 near B=[100] showing prismatic loops; and deformed at 77K: (e) long straight dislocations, g=01 1 near B=[100]. Strain rate = 10" 4

PAGE 91

84 Figure 4.26 — continued

PAGE 92

85

PAGE 93

86 Figure 4.27 BFTEM images showing dislocation substructure in the [557] specimens deformed at 473K at a strain rate of 10" 3 s" 1 (a, b) (elongation=18%) and lO'V (c, d) (elongation=19%): (a) slip plane view; (b) slip plane edge-on; (c) slip plane view and (d) slip plane edge-on.

PAGE 94

87 Figure 4.27 -continued

PAGE 95

88 Figure 4.28 Dislocations in [557]S(1 10) specimen tested at 673K: (a) dislocation substructure, (b) subboundary, and (c) dragging of the dislocations by jogs. TEM foil cut parallel to the Face from the necked region in the gage of tensile specimen. Strain rate = 10 s '. g=011 near B=[ 111].

PAGE 96

89 Figure 4.28 continued

PAGE 97

90 Figure 4.29 Dislocation substructure in the [123] oriented specimens after 3% deformation (strain rate = 10" s" 1 ) at 723K showing (a) bands and (b) cross slip segments out of the (110) primary slip plane (arrow points). B near [001].

PAGE 98

91 Figure 4.30 Dislocation sub-boundary and networks in the [557] oriented specimens deformed at 1073K and a strain rate of (a, b) 10" 4 s" 1 , and (c) 10" 2 s" 1 .

PAGE 99

92 Figure 4.30 -continued

PAGE 100

93 necking and higher tensile elongation than that tested at the lower strain rate (10' 4 s" 1 ) (Figure 4.5). In general, the temperature affects the deformation of soft-oriented NiAl single crystals via the gradual increase in the activity of both cross slip and climb mechanisms with increasing temperature. Likewise, the strain rate affects deformation in a similar manner and the combination of temperature and strain rate dictates the relative WHR and recovery rates of NiAl single crystals. No fundamental change in the slip behavior is observed over the temperature range studied. 4.5 Slip Behavior of Hard-Oriented NiAl Single Crystals In order to identify the slip behavior in <100>-oriented NiAl and NiAl-Si single crystals, TEM foils were cut from tensile specimens after 2-5% tensile deformation at 473K, 523K, 573K, 723K, and 873K. The foils were cut parallel to the (100) (NiAl) and (1 10) (NiAl-Si) planes from specimens whose tensile axis was [001]. 4.5.1 Dislocations in the specimens tested below 573K Only cube <001> dislocations were observed in specimens deformed below 573K. The deformation appears quite heterogeneous since in some TEM foils, no dislocations were observed while in others, dislocation substructures such as that shown in Figure 4.31 were observed. This particular micrograph was taken from a specimen tested at 523K to -2% elongation before fracture. Standard g.b and trace analyses indicated that these dislocations are all [010] dislocations mostly lying on the (100) plane. There are many prismatic loops elongated along [001], similar to what is observed in the [01 1] oriented RT specimens. However, since there is no resolved shear stress for <001> slip when the specimen is oriented accurately along <100> directions, activation of cube slip is not expected. One possible mechanism of producing cube slip in <100> oriented NiAl is by

PAGE 101

94 kinking; this leads to local activation of <001> slip as pointed out by Fraser et al. (65). Although no kinking was observed in the specimen gage before the TEM foils were cut, it is possible that microscale kinking occurred. It is also possible that the [010] dislocations were introduced during TEM sample preparation. However, since the sample preparation in this work was carefully controlled and no such dislocation structures were observed in unstrained specimens or those tested at other temperatures similarly prepared, it is unlikely that these dislocations are introduced during sample preparation. Consequently, although these dislocations are not believed to represent the slip behavior in <100> oriented NiAl during uniform deformation, they must be some sort of by-product of the actual mechanisms involved in the deformation. 4.5.2 Slip Behavior at 573K At or above 573 K, <100> oriented NiAl has tensile elongations typically greater than 20% so that a uniform activity of some slip system is expected. Figure 4.32 shows the dislocation substructure in hard crystals deformed at 573 K and it can be seen that the dislocations display a well defined distribution. Except for considerable debris and very small loops, most dislocations belong to one of the three types marked as A, B and C in the figures. An interesting feature is that the projection of A dislocations along B=[100] lies in the direction of [Oil] while that of B dislocations in the direction of [OlT]. The majority of dislocations are of type A and B. However, there are some C type dislocations which normally do not lie along either [01 1] or [OlT] directions, g.b analysis shows that A and B dislocations are both in contrast with g=002, g=020 while A dislocations are out of contrast with g=0lT and B out of contrast with g=011. This implies that A and B are either <011> or <111> dislocations. However, tilting the specimen 45° to the [1 10] and [HO] zone verifies that A dislocations have a [01 1] Burgers vector while B dislocations have a [OlT] Burgers vector, as shown in Figure 4.32(e). It is

PAGE 102

95 further identified that the slip systems are [01 l](0lT) and [0lT](01 1). Figure 4.32(a) shows both slip planes edge on for both A and B dislocations which indicates that the slip is confined to the slip plane. The C dislocations have <010> Burger vectors. It is also interesting to note that both A and B dislocations show zigzag line features in the [110] and [llO] zone TEM images (Figure 4.32(d-e)). Since A and B dislocations are confined to their slip plane as shown in Figure 4.32(a), these zigzag features suggest that the these <01 1> dislocations have well-defined line directions in the slip plane. A detailed trace analysis (Figure 4.33) revealed that the dislocation lines of <1 10> dislocations are zigzagged in the slip plane with their segments parallel to one of the two <1 1 1> directions contained in the slip plane. Specifically, a [Oil] dislocation will have segments lying along [1 1 1] and [Tl 1] alternatively as shown in Figure 4.33(d). This geometry indicates that (1) the motion of <011> dislocations is quite retarded so they do not bow out under stress, (2) the <011> dislocations undergo rapid relaxation upon specimen unloading, or (3) the <1 1 1> line segments have lowest mobility or highest Peierls stress. The high density of <011> dislocations indicates that <011> dislocations have reasonable mobility so the first mechanism seems unlikely. The relaxation of the dislocations may not explain the <111> line direction either, since the modeling efforts show that the lowest line energies correspond to the screw orientations, i.e., u parallel to <011> (51, 53). However, the theoretical calculations to date have not taken into consideration the configuration of the <110> dislocation core structure, which have been observed by HRTEM and WBDF to be either dissociated or decomposed (63). Therefore, the results from the theoretical modeling may not be reliable. In addition, there has been no theoretical calculations regarding the Peierls stress for dislocation line configurations other than those with screw or edge character. Significantly, the <1 1 1> line directions on {01 1 } planes pass through alternative Al and Ni atom columns as shown in Figure 4.34 which should make its core structure very different from that when the dislocation lies along the

PAGE 103

96 <001> (edge) or <01 1> (screw) directions. As a result, it seems reasonable to expect the mobility of <1 1 1> line segments to be lower than the other segments. The alternative [Til] and [1 1 1] line segments of <01 1> dislocations may result from the kink motion of <1 1 1> segments. Nevertheless, the preferred line direction of <011> dislocations in this study indicates that a better understanding of both the mobilities and line energies of <011> dislocations is needed. Unlike <001> slip in soft-oriented NiAl, <011> slip is relatively uniform as indicated by the lack of dislocation bands in the latter. There are, however, some regions where narrow <011> dislocation bands containing a high density of debris inside them were observed although the debris in this case is not formed via mechanisms similar to those described for the soft-oriented crystals. This is most likely due to very limited cross slip of <01 1> dislocations. These debris and small loops are actually formed via interaction between the <01 1> dislocations as will be shown later in this section. It is also of interest to note that another set of <101>{ 101 } slip systems were also activated in the specimen (Figure 4.35). It is well known that in NaCl crystals and in some other materials which deform exclusively by <01 1> slip, activation of one pair of <01 1> slip systems will result in latent hardening and inhibit another pair of <01 1> slip systems. In NiAl, if the initial orientation of the specimen is off the accurate [001] direction, the Schmid factors for the <01 1> slip pairs is different and the activation of one of the <01 1> slip pairs is favored; this could result in latent hardening. However, the observation of both the <01 1> pairs in this specimen suggests that either the latent hardening effect is not obvious in NiAl, or the specimen orientation is very accurate. In order to determine if Si has any effect on the slip behavior at this temperature, the slip systems in <100> oriented NiAl-Si specimens after tensile testing at 573K were also investigated via g.b and trace analyses. As in the case of binary NiAl, <011> is the dominating slip vector in the <100> oriented NiAl-Si single crystals, as shown in Figure

PAGE 104

97 4.36a and b. Figure 4.36b shows that there are no <1 1 1> dislocations. Figure 4.36b also shows that only the [01 1](01T) and [0l!](01 1) slip systems are operating. Unlike in the case of <001> oriented NiAl deformed at the same temperature, there is no evidence of the [101](10l) and [10l](101) systems in the NiAl-Si specimen, which indicates the effect of latent hardening. Although the slip systems are the same in both the NiAl and NiAl-Si specimens, the line directions of these <011> dislocations in the NiAl-Si specimen are less well defined (Figure 4.36c). The dislocation lines in the Si-doped materials also more tend to bow out more both in and out of the slip plane. These features indicate that the <01 1> dislocations in Si-doped NiAl are more mobile and cross slip more readily. There is also a higher density of debris and small loops in the region where the <011> dislocations are clustered (Figure 4.36c). At first glance, these debris and loops could be confused with the edge-on images of other <01 1> slip systems. However, careful analysis showed that they are distinguishable. In one such [Oil] dislocation band, the debris and small loops were identified to have either a [010] or [001] Burgers vector. This may indicate that these debris are products of the following dislocation reactions: a[011]+a[0lT]=2a[010] a[011]+a[0Tl]=2a[001] and occur when [01 1] and [OlT] dislocations intersect. This also indicates that the <01 1> dislocations are reasonably mobile at this temperature. This type of reaction has been observed in the specimens, as shown in the WBDF images of the NiAl deformed at 573K (Figure 4.37). Although, decomposition of [011] dislocations can also result in these same <001> segments (a[01 l]=a[010]+a[001]), it seems highly unlikely that this decomposition would produce the debris or small loops observed.

PAGE 105

98 4.5.3 Slip behavior at temperatures above 573K In the <001> NiAl specimens deformed at 723 K, the dislocation substructure mostly consists of cube dislocations (Figure 4.38). However, a small fraction of <01 1> dislocations can still be observed (see arrows in Figure 4.38a and Figure 4.38c) and slip on {011} planes. The dislocation loops are larger than those in the specimens deformed at 573K and are identified as <001> loops. The dislocation density is about the same as that at 573K even though the majority of the dislocations are the <001> type. The dislocation substructure in the [001] -oriented NiAl specimen deformed at 873K appears more like a recovery structure with many <001> loops and a relatively low dislocation density (Figure 4.39). The majority of the dislocations have been identified via g.b analysis to be <001> dislocations. A very small percentage of <011> dislocations exists and only as segments which are always sandwiched by <001> segments (see arrow in Figure 4.39); this suggests a possible decomposition of <01 1> dislocations into <001> dislocations. Among the dislocations identified, less than 10% are <01 1> type, about 45% are [001] and the remaining 45% are [100] and [010] which also suggest the <011> to <001> decomposition. The change of the relative densities of the dislocations with testing temperature is shown in Figure 4.40 and the shift from the dominance of <01 1> to <001> dislocations is obvious. The dislocation substructures in the [001] NiAl-Si specimens deformed at 723K are generally similar to those in binary NiAl specimens (see Figure 4.41). However, the NiAlSi specimens appear to have a higher percentage of <01 1> dislocations and a higher density of <001> loops. It also appears that the operation of <011> slip systems in NiAl-Si extends to a higher temperature than in NiAl single crystals. This may suggest that Si stabilizes the <011> dislocations against their decomposition into <001> dislocations. Considering that the transition in the temperature dependence of the yield stress of the

PAGE 106

99 Figure 4.31 Dislocations in the 523K deformed <100> NiAl. (a) Figure 4.32 Dislocation substructure in the [001] oriented NiAl single crystals after 2 5% strain at 573K: (a) g=020 near B=[ 100], (b) g=01 1 near B=[_100], (c) g=0TT near B=[100], (d) g=l 10 near B=[l 10], and (e) g=l 10 near B=[l 10]. TEM foil cut parallel to [100]. TA = tensile axis.

PAGE 107

Figure 4.32 -continued

PAGE 108

Figure 4.32 continued

PAGE 109

102 Figure 4.33 The zigzag line configuration of <01 1> dislocations in the <100> oriented specimen deformed at 573K: (a-c) Higher magnification images of type A and B dislocations taken from the dash line marked areas of Figure 4.32, (d) illustration of b=[01 1] dislocation line directions.

PAGE 110

Figure 4.33 — continued

PAGE 111

104 Figure 4.34 Illustration of the configuration of a [101] dislocation having <1 1 1> line directions. The dislocation line runs alternatively through unlike atom columns, unlike that in the case of pure screw or pure edge, dislocation line runs through only like atom columns.

PAGE 112

105 (b) Figure 4.35 The [01 1] and [101] dislocation bands in the <100> oriented NiAl specimen after deformed -2.5% at 573K for Foil normal_parallel [100]: (a) g=020 and B near [100] showing [101] bands out of contrast, (b) g=01 1 showing [101] band in contrast.

PAGE 113

Figure 4.36 Dislocations in [001] oriented NiAl-Si single crystal after deformed -4% at 573K: (a) g=002 near B=[100],_(b) g=020 near B=[100] showing all the dislocations out of contrast, (c) g=002 near B=[l 10] and (d) g=020 near B=[100]. TEM foil normal parallel to [110].

PAGE 114

107 Figure 4.36 -continued

PAGE 115

108 (b) Figure 4.37 Weak beam dark field (WBDF) TEM image of dislocation substructure in the [001] NiAl specimen after testing at 573K. Foil normal parallel to [100]. The short arrow shows where two <01 1> dislocations come across each other and form two <001> segments. Arrows point to (a) [011] dislocation with [001] segments, and (b) [011] dislocation with the [00 1 ] segments.

PAGE 116

109 Figure 4.38 Dislocation substructure in the [001] oriented NiAl single crystal after deforming 723K deformed about 3%. TEM foil normal parallel to [100].

PAGE 117

110 Figure 4.39 Dislocation substructure in the [001] -oriented NiAl single crystal after deforming at 873K about 3%: (a)_g=002 near B=[100], (b) g=01 1 near B=[100] (c) g=020 near B=[100], and (d) g=l 10 near B=[l 10]. Arrows point to a [101] segmer Foil normal [100].

PAGE 118

Ill 100 0 J — I — — 1 4 — J 373 573 773 . 973 Temperature (K) Figure 4.40 Relative dislocation density in [001] oriented NiAl single crystals as function of tensile testing temperature. All the specimens are strained to 3-5% tensile elongation. Figure 4.41 Dislocation substructure in the [001] oriented NiAl-Si single crystal after deforming at 723K for 4.5%: (a) g=01 1 near B=[100] and (b) g=01 1 near B=[100]. Foil normal [100].

PAGE 119

112 NiAl-Si specimens (Figure 4.7 and Figure 4.9) also occurred at higher temperatures than in binary NiAl and that 723K is the midway through this transition for the NiAl-Si alloy but about the end for binary NiAl specimens, it may suggest that the sudden drop in the yield stress is due to the presence of large amounts of <001> dislocations. However, the instability of <011> dislocations needs to be verified via further experiments. Since the BDTTs of both the NiAl and NiAl-Si specimens, as derived from the temperature dependence of the tensile elongation (Figure 4.8), appears to be between 523 and 573K, the slip behavior observed in the above specimens indicates that the BDT was due to the activation of <01 1> slip, and not due to the climb of <001> dislocations as suggested by other workers (87). Since no dislocations other than <001> type were observed in <001>oriented tensile specimens tested below 573 K, it was not possible to conclude if there is a <1 1 1> to <1 10> transition in this material as reported in previous studies (88). 4.6 Slip Traces and Fracture Surfaces 4.6. 1 Soft Orientation Whenever possible, slip traces on the surface of the tested specimens were also analyzed. For the specimens deformed at temperatures above 473K, the surface slip traces were not readily distinguishable. At lower temperatures, the surface slip traces were rare and very non-uniformly distributed even on the specimens deformed over 10% at RT. In those instances where the surface slip traces or marks were observed, they were usually only visible in limited areas close to the fracture surface where the stress states may have been more complicated. Nevertheless, these slip traces may still provide additional information about the slip processes occurring in the material. Figure 4.42 shows respectively the surface marks on a [557]F(1 10) specimen tested at RT and two [557]F(1 10) specimens deformed at 473K at different strain rates. The slip

PAGE 120

113 traces on the RT specimen are essentially parallel to [1 10] on the (7 7 10) surface consistent with [001](Tl0) slip traces (Figure 4.42d). These marks tend to be quite wavy consistent with considerable cross slip. It should be pointed out that numerous microcracks were observed at the larger slip offsets in agreement with the results of a recent study by Ebrahimi et al. (90); this suggests that the cracks may actually initiate at the edge or intersection of the observed dislocation bands where stress concentrations are usually high. The surface marks on the specimens deformed at 473K using strain rates of 10" 4 s" 1 and 10" 3 s 1 are less well defined than those specimens deformed at RT. Specifically, the surface marks on the (7 7 10) gage surface of the specimen tested at 10" 4 s" 1 appear more wavy and broader than those on the specimen tested at 10" 3 s" 1 , the latter being more welldefined and consistent with traces of surface steps lying along the expected [1 10] directions as in the RT specimens. It is also noted that there were no apparent surface microcracks on the surface of the 10" 4 s' 1 specimen, whereas microcracks were observed on the specimen tested at the higher strain rate (Figure 4.42c). These differences suggest that, at the strain rate of lO'V, the slip may be more uniform and that the stress concentration does not build up as much as in the specimen tested at the higher rate. This is consistent with the TEM analysis described in a previous section. The fracture surfaces of soft crystals tested at temperatures from RT to 600K are shown in Figure 4.43. The fracture surface of a typical specimen tested at room temperature shows river-like brittle fractures feature which indicates fast propagation after crack initiation. Even at 573K, where high elongations could be achieved due to the balance between the recovery and the strain hardening of the specimen, the fracture surfaces still appear brittle although the finer features indicates slower crack propagation than at lower temperatures, or more complex crack path. At 873K and lOV, the specimens neck down to chisel points (Figure 4.43c); this is related to the faster recovery processes at this temperature. It is noted that the deformation at 873K was still dominated by the [001](Tl0)

PAGE 121

114 system as indicated by the shape change during specimen necking, i.e., by the shearing on (TlO) planes. Figure 4.43c also shows that, in contrast to the necking at 873K, the crosssectional area of the specimen deformed at 573K decreased more uniformly throughout the gage length; this uniform reduction in the cross section area is carried out by the spreading of the propagation bands as described elsewhere (112). 4.6.2 Hard Orientation The gage and fracture surfaces of the binary and NiAl-0.3Si crystals tested between 573K and 873K were investigated in order to understand the tensile elongation behavior more clearly (Figure 4.8). It was found that the slip traces are even less common than those in the soft orientation and could not be used to perform reliable trace analyses. As with the soft-oriented crystals, the fracture surfaces of the hard orientation specimens (Figure 4.44) exhibit a brittle fracture nature. In the case of specimens deformed at 573K, it is also noted that on both the fracture surface and the gage surfaces, many large secondary cracks had penetrated deep into the specimen (Figure 4.44a); this may be due to the higher flow stress at 573 K since much fewer secondary cracks were observed on the surfaces of the specimens tested at 723K where the flow stress is considerably lower (536 MPa vs. 867 MPa) (Figure 4.44b). The fracture of the specimens tested at 723K was also brittle as shown in Figure 4.44b, although even higher tensile elongations (-50% vs. -20%) were observed. Finally, at 873K, the specimens neck down to chisel points (Figure 4.44c) similar to that observed in the soft-oriented crystals deformed at the same temperature; this indicates a rapid recovery process and relatively easy dislocation motion. As noted above, the dislocation density in binary crystals deformed at this temperature is quite low (Figure 4.39), consistent with this rapid recovery. Taking into consideration that 873K is also the temperature where both soft and hard crystals exhibit similar yield stresses (Figure 4. 10), it appears that, at this temperature, the deformation of both soft and hard crystals is controlled

PAGE 122

1 15 Figure 4.42 SEM images of the surface slip traces of [557] oriented NiAl crystals: (a) after 8% elongation at RT (strain rate = 10" 4 s" 1 ) showing the microcracks (arrows) associated with these slip traces, (b) after 19% elongation at 473K (strain rate = 10" 4 s" 1 ) showing coarse and wavy surface marks, (c) after 18% elongation at 473K (strain rate = 10" 3 s'*). (d) Schematic illustration showing the intersection of the slip plane and gage surfaces.

PAGE 123

116 Figure 4.42 continued

PAGE 124

117 Figure 4.43 Fracture surface of: (a) a [557]F(1 10) specimen tested at RT, (b) a [123] oriented NiAl single crystal specimen tested at 573 K after 100% elongation and (c) the specimens oriented close to [1 1 1] and deformed at 573K and 873K, respectively Strain rates = 10 s for all the specimens.

PAGE 125

118 LO mm Figure 4.43 — continued

PAGE 126

119 Figure 4.44 Fracture surfaces of the [001] oriented NiAl-0.3Si specimens fractured at (a) 573K, (b) 723K and (c) 873K at a strain rate of 10" 4 s\

PAGE 127

120 Figure 4.44 — continued

PAGE 128

121 by the rate of recovery. Since the climb of <001> dislocations is clearly important to the deformation. This is indeed consistent with the high percentage of <001> dislocations present in the hard orientation crystals. However, the high flow stresses of hard-oriented crystals at 573K and 723K indicate that the glide of <011> dislocations is much more sluggish than that of the cube <001> dislocations in the soft orientation crystals, which is also consistent with the fact that the movement of <01 1> dislocations is confined to their slip plane. Therefore, the sudden increase in tensile ductility at 573K in hard orientation crystals appears to be related to changes in the mobility of <01 1> dislocations and, unlike in the deformation of soft orientation crystals, the tensile elongations of <001> crystals reach their highest values when the recovery (i.e., climb of <001> dislocations) become sufficiently high since the dislocation reactions such as <0lT>+<001>=<010> may be required to avoid high stress and strain concentrations due to the lower mobilities of the <01 1> dislocations. In general, the fracture surface analyses indicate that the fracture of NiAl single crystals is always brittle consistent with the fact that maximally only 3 independent systems exist in the soft-oriented crystals and only 2 in the hard-oriented crystals. 4.7 General Discussion 4.7. 1 Deformation of NiAl Single Crystals It is now obvious that the deformation of NiAl single crystals is governed by: (1) slip of <001> dislocations which can only result in at most three independent slip systems when oriented along soft directions and (2) slip of <01 1> dislocations at and above 573K when oriented along the hard <001> direction. The generation and glide of <001> dislocations also appears to occur readily.

PAGE 129

122 As for the deformation of soft-oriented NiAl single crystals, it has long been argued that the CRSS of NiAl is both orientation dependent and slip plane dependent, and that cube slip on the MRSS plane occurs. However, the results of this study indicate that there is no significant difference between the CRSS of <001> dislocations on the {100} and {110} slip planes. As noted above, the primary slip plane in NiAl can be determined unambiguously via a very simple TEM technique which utilizes residual contrast in BFTEM images taken with the beam direction parallel to the Burgers vector, as illustrated in Figure 4.45. This is possible because the traces of debris and prismatic loops created by dislocation glide are always aligned with the edge direction in the slip plane, and the traces of dislocation bands are the primary slip planes. Using this technique, it was determined that the primary slip planes were always either { 1 10} or {010}. Although the results of this study do not exclude the possibility Of {hkO} planes being micro-scale cross slip planes, as needed for the proposed double cross slip mechanism, it appears that {hkO} planes are not the primary slip planes. The {hkO} surface slip traces reported by other investigators (25) could be due to composite cross slip between the active { 1 10} and {010} planes, or orthogonal {110} planes as suggested by Ball and Smallman (2). In fact, a careful analysis of the deviation angles reported by Takasugi et al. indicates a temperature dependence which could be extrapolated to the {110} slip plane at OK for the <123>oriented single crystals tested in their study, as shown in Figure 4.46. This phenomenon can be readily explained by a gradual decrease in the activity of <001> cross slip between the primary {110} plane and other planes. Indeed, it is observed in this study that the cross-slip activity decreases with decreasing temperature and increasing strain rate. Therefore, it is believed that the {hkO} surface slip traces in their study are not the results of MRSS slip but rather an indication of cross slip. Cross slip planes other than {010} and { 1 10} are also deemed less likely due to the nature of the B2 structure, and the absence of any evidence of slip on the MRSS planes.

PAGE 130

123 Nevertheless, in soft NiAl single crystals, the screw components of <001> dislocations cross slip readily whenever they encounter obstacles and this accounts for the formation and broadening of the dislocation bands. Therefore, the deformation of soft NiAl single crystals should be easier if there are few obstacles in the material (e.g., high purity NiAl at the beginning of deformation), or the obstacles could be circumvented or annihilated given sufficient time or thermal activation (e.g., via cross slip and/or climb). The situation in the deformation of hard-oriented crystals is different from that of soft-oriented specimens. In this case, the Schmid factor for <001> slip is zero and the <1 1 1> and <01 1> slip systems are not readily activated below 573 K. Activation of the <01 1> slip systems is clearly responsible for the deformation of hard-oriented crystals at temperatures above 573K. However, the results of this study indicate that the movement of <01 1> dislocations is sluggish and the increase in tensile ductility of hard oriented crystals at higher temperatures is at least partly related to the climb of <001> dislocations as will be discussed later in section 4.7.3. 4.7.2 Tensile Ductility of Soft Orientation It has been observed that NiAl single crystals undergo a rapid increase in either tensile ductility or fracture toughness with increasing temperature and this increase has often been referred to as the BDT. However, unlike the BDTT in most traditional metals such as Mo where it occurs at a much lower homologous temperature (<0.1T m ), the BDTT in soft orientation NiAl is reported to be around 0.25T m . In most traditional metals, the BDTT is either dislocation mobility or nucleation controlled whereas the generation and motion of <001> dislocations in NiAl has been found to occur readily at temperatures below the BDTT. Unlike the cases of Mo or Ti, there are not sufficient independent slip systems in NiAl to satisfy the Von Mises criterion. This means that, even though dislocation motion occurs easily, the lack of a greater number of slip systems leads to the

PAGE 131

124 development of stress concentrations during plastic deformation. The density of mobile dislocations in NiAl is also typically high which makes the mechanisms responsible for the BDTT of brittle materials such as Si and alumina unlikely for soft-oriented NiAl. All these characteristics makes the definition of the BDTT in NiAl somewhat controversial even though both tensile elongation and fracture toughness display definable rapid increases with temperature (66, 90). Furthermore, the tensile ductility of high purity crystals used in this study exhibit a much more gradual increase with temperature (see Section 4.1) until an abrupt increase occurs in the intermediate temperature range due to a balance between the recovery and strain hardening rates. Therefore, it seems that the high tensile ductility and the "BDTT" of NiAl should be distinguished from each other. Due to the gradual decrease in tensile elongation with decreasing temperature, it may be more appropriate to describe the tensile ductility of NiAl based on its gradual change rather than by a sudden transition defined by a BDTT. The results of this study suggest that the ductility of soft crystals may be controlled by a combination of cross slip processes and diffusion assisted processes, as described in detail below and illustrated in Figure 4.47: (1) At high strain rates or low temperatures, very limited ductility is expected. The dislocations in these situations undergo less cross slip, and have relatively lower mobilities, as indicated by the less bowed dislocations in the specimens deformed at 77K and 120K. The drag on the moving screw dislocations exerted by the jogs that were created by the double cross slip mechanism is also more severe in this case due to both the drifting of the jogs and the pinching-off of the debris and prismatic loops requires diffusion-assisted climb locally. All of these factors lead to a higher probability for local stress concentrations to build up at obstacles or dislocation bands. Since the CRSS increase with decreasing temperature and/or increasing strain rate, microcracking at stress concentrations and reduced ductility is expected.

PAGE 132

125 (2) At sufficiently slow strain rates or high temperatures, the thermally activated cross slip and the easy drifting of the jogs or the climb of dislocations can readily reduce the local stress and strain concentrations at dislocation bands or other obstacles. The single crystals will show substantial elongation in this case. The amount of tensile ductility at this stage is usually controlled by the extent of plastic instability (necking) in the specimen. When a certain balance between the strain hardening and recovery processes is reached, the plastic instability is resisted and the specimens can be elongated enormous amounts (>300%). (3) At intermediate strain rates or temperatures, the glide of dislocations leads to the formation of localized bands and cross slip results in debris, prismatic loops as well as new dislocation sources that actually broaden the bands. The ductility is controlled by a balance between the reduction rate of local stress concentrations via cross slip and the rate of development of stress concentrations as deformation continues. Ultimately, microcracks form and propagate when the local stress reaches the cleavage value at the dislocation bands or other stress concentration sites. Therefore, statistically, ductility gradually increases as the temperature increases or strain rate decreases, due to a shifting in this balance. For the cases of lower purity NiAl, water quenched NiAl and off-stoichiometric or alloyed crystals, the higher CRSS and lower dislocation mobility result in higher stresses that tend to reach the critical values for microcrack nucleation more quickly; this is reflected in the lower tensile ductilities observed. This particular mechanism is supported by the fact that the cross slip activity increases with temperature in this material as discussed in previous sections. It is also noted that limited tensile elongations (up to 6%) can be obtained for soft-oriented high purity crystals tested at temperatures as low as 120K at a strain rate of 10" s" 1 . This suggests that, at the beginning of deformation, the slip of dislocations is not impeded significantly. Therefore, except for the extreme cases at very low temperatures or where

PAGE 133

126 very high strain rates are applied, the ductility of soft orientation crystals depends on how the local strain and stress concentrations develop. It has been observed (90) that the initial work hardening rate was quite high in the <110> oriented NiAl single crystals tested between RT and 473K and then decreased significantly at higher strain levels; this may be explained as being due to an enhanced cross slip activity at higher strains. It was also found by the same authors that the strain hardening rate was very sensitive to the strain rate, and sensitive to temperature changes over a certain range of strain rates. This also suggests a thermally activated process for the development of stress or strain concentrations. As discussed above, the effects of temperature and strain rate on the ductility can be correlated with their effects on both the thermally-assisted cross slip and on the movement of jogged dislocations. The reduction of local stress concentrations by double cross slip is more effective when the spacing (h) between the initial primary slip plane and the parallel primary slip plane is larger than some critical value \, which is the condition for forming new dislocation sources rather than edge dipoles (Figure 4.48). This process consists of forming a screw segment, ejection of the screw segment onto the cross slip plane for a distance larger than hg and then re-ejection back onto the parallel primary slip plane and clearly, is thermally assisted (113). In NiAl, the screw dislocations are not dissociated so the thermal activation energy for forming screw segments may be very small. However, a higher thermal activation energy may be needed for dislocations to slip onto the cross slip plane for a distance larger than h,,. Assuming that the rate of such double cross slip events can be expressed as R=R 0 exp(-Q cs /KT), while is the activation energy and R,, a parameter representing other effects on cross slip, e.g., the tensile axis, then a critical strain rate that can be accommodated by such cross slip process without statistically increasing the local stress concentrations should be proportional to R. For h-dio , it is necessary for the jogs on a jogged dislocation to move with the dislocation, or easily pinch off prismatic loops to allow the dislocation to move away. The

PAGE 134

127 effect of temperature or strain rate on the movement of the jogged dislocations can be estimated as follows. A dislocation line will be pinned by sessile jogs after cross slip, assuming the jogs move with the dislocation which is the case of very small to intermediate jogs (h
PAGE 135

128 This strain rate represents a rough estimate of the critical strain rate at which the velocity of the jogged dislocations can accommodate the deformation of the specimen. Apparently, the above-mentioned processes may contribute differently to the deformation depending on temperature. It can be expected that at lower temperatures where diffusion is slow, the cross slip process may dominate whereas at higher temperatures, the drifting of the jogs may be more important. Although other factors such as the CRSS should also affect the probability of crack initiation and propagation, the temperature and strain rate dependence of the tensile ductility can be explained with the two processes just discussed. If the loading strain rate is much higher than the critical strain rate determined by these processes, cross slip may not effectively reduce the local stress concentrations and a low ductility is expected. If the loading strain rate is much lower than the critical strain rate, the local stress concentrations can be released effectively via cross slip and therefore, a higher tensile is expected. These relationships indicate that the tensile ductility of soft-oriented NiAl should be both strain rate and orientation dependent in agreement with the results of Ebrahimi et al. and Takasugi et al. (6, 90). Ebrahimi et al. reported that the BDTT of <1 10> oriented NiAl single crystals shifted from 473 K to below RT when the displacement rate was decreased from 10" 2 mm/sec to 2xl0" 5 mm/sec. They also reported an activation energy ranging between 50 and 100 KJ/mol depending on the applied stress or stress intensity level; this activation energy is substantially lower than the self diffusion energy reported for NiAl (about 200-300 KJ/mol) (12) and seems to support the double cross slip mechanism. However, since the Q L represents the activation energy for local diffusion such as pipe diffusion along dislocation cores in highly strained regions, it can be much lower than the activation energy for bulk self-diffusion (1 14, 1 15). Nevertheless, both the cross slip of screw segments and the motion of jogged dislocations are thermally activated, consistent with the strain rate and temperature dependence of the tensile ductility in NiAl crystals.

PAGE 136

129 It should be pointed out that, although the tensile ductility is not suitable for a direct representation of the brittle-to-ductile transition, the results of this study and the mechanism of tensile ductility revealed by the current results are in good agreement with the conclusions generated from a fracture toughness study of the BDTT by Ebrahimi et al. (90) as it has been found in their study that the thermally activated cross-slip reduces the slip localization tendency in NiAl and, therefore, enhanced cross-slip of dislocations results in the improvement in the brittle-to-ductile transition. 4.7.3 Tensile Ductility of Hard Orientation NiAl Crystals Field et al. (28) reported that hard-oriented NiAl single crystals have no tensile ductility at temperatures below 589K and the results of this study are consistent with their results. At or above 573K, a sufficient amount of tensile plasticity could be obtained similar to all other previous reports on hard orientation crystals. The sudden increase in tensile ductility of the <001> crystals around 573K is not well understood. One of the mechanisms suggested for this sudden increase, referred to as the BDTT for the hard orientation crystals, is <001> climb (47, 87) due to the similar strain rate sensitivity between hard crystals and polycrystals. However, it was found in this study that in the 573K deformed specimens, where more than 20% elongation was achieved, only <01 1>{01 1 } slip systems were active. This indicates that the activation of <01 1> slip is the cause for the onset of tensile ductility in the hard crystals. On the other hand, although <001> dislocations were observed in a large quantity in the specimens deformed at 723 K, it is unlikely that the climb of <001> dislocations is the main contribution to the deformation unless a sufficiently high temperature and/or low strain rate is used. Furthermore, unlike in the case of soft crystals, the transition in the ductility of hard crystals happens over a narrow temperature range which resembles that observed in Si which occurs over a temperature range of several degrees (116). Therefore, a dislocation mobility

PAGE 137

130 controlled mechanism similar to that in Si may be responsible for the BDTT. This then poses the question as to why the mobility of <01 1> dislocations increases with temperature and what mechanisms control the mobility of <01 1> dislocations around 573 K. It is found that the ductility transition in <001> oriented NiAl is very sensitive to strain rate (11). Other experimental results also indicate that this transition is sensitive to deviations from stoichiometry (88) and alloying additions (87). In addition, the results of a HRTEM study of the core structure of <01 1> edge dislocations by Mills and Miracle (63) indicate that the core of <01 1> dislocations is either climb-dissociated or decomposed into two adjacent <001> cores. Decomposition of <011> dislocations into <001> dislocations at high temperatures has also been indicated by the results of this study and previous studies (28). It is therefore a reasonable assumption that the core structure of <011> dislocations causes the abrupt transition in its mobility and thus the tensile ductility. Mills et al. explained the low mobility of <011> dislocations using diffusion assisted glide mechanisms since the glide of <011> dislocations can be carried out by cooperative localized climb of two <001> edge dislocations as shown in Figure 4.49. Their mechanisms are consistent with the fact that the <01 1> dislocations appear to be unstable at higher temperatures where a high density of <001> dislocations are observed after deformation. However, it is difficult to understand why the cooperative climb should occur since, if the decomposition is energetically favored, there is no reason to expect the <001> dislocations to stay together once decomposed, especially when the loading stress will favor climb of the <001> dislocations with their Burgers vectors parallel to the loading axis. In this study, <01 1> dislocations have been observed to lie on the slip plane with a skew configuration which could be thought of as having many large kinks in <111> directions. Therefore, the glide of <01 1> dislocations may be carried out by kink motion as discussed in Section 4.5. It appears that such kink motion requires thermal activation

PAGE 138

131 resulting in <011> dislocations that only become sufficiently mobile to allow the deformation around 573 K for a strain rates of 10"V. Changes of the core configuration of <011> dislocations should affect their mobility, which could explain the composition dependence of the transition temperature as it is reported that the elastic constants of NiAl are sensitive to composition (Figure 2.3). It was also found in this study that, at high temperatures, the ductility of hard-oriented crystals may be due to the recovery process (i.e., <001> climb) which lowers the stress concentrations that <01 1> glide is not able to effect. Therefore, although the onset of tensile ductility is due to the activation of <01 1> slip, the ductilities at higher temperatures may not be related to the mobility of <01 1> dislocations as much as the increased role of recovery due to <001> climb. It should be pointed out that the activation energy for deformation at 700K reported by Pascoe et al. (1 17) showed a value close to that for self-diffusion in NiAl although the authors also reported much higher activation energies at higher temperatures which are not readily explained. Lahrman et al. (11) also studied the strain rate sensitivity of <001> oriented NiAl. Based on the reported 50K increase in the transition temperature with an increase of strain rate from 8.3xl0" 5 to 8.3xl0" 3 s"\ the activation energy was estimated to be about 270 KJ/mol which is also close to the self diffusion energy in this material. While these data seem to support the diffusion-assisted glide mechanisms, it is noted that if the recovery process is the key to the tensile ductility of hard crystals at and above 700K, the activation energy should be the same for either of these mechanisms. Actually reducing the diffusion rate should result in a slower recovery process, thus reducing the tensile ductility. This expectation is consistent with the results of a few studies where the diffusivity in NiAl was reduced greatly via alloying and, accordingly, the transition temperature was shifted to higher temperatures (87). Nevertheless, the onset of tensile ductility of <001> oriented NiAl is found at the temperature where <011> dislocations become sufficiently mobile at a temperature and

PAGE 139

132 strain rate where substantial amounts of climb of <001> dislocations is unlikely. Therefore, the sudden increase in the ductility is attributed to the activation of <011> slip systems. 4.7.4 Potential of NiAl Single Crystals as High Temperature Structural Material Based on the discussion in section 4.7.1 through 4.7.3, it is now obvious that the poor low temperature toughness and ductility of NiAl single crystals is an intrinsic property in that there is an insufficient number of independent slip systems to release the stress concentrations that develop during inhomogeneous local slip. Extrinsic factors such as thermal vacancies (heat treatment), impurities and alloying additions affect the ductility by retarding dislocation movement (i.e., raising the CRSS) thereby exacerbating the stress concentrations. Consequently, any efforts to improve the low strength of soft orientation NiAl via approaches of retarding dislocation slip without increasing the number of independent slip systems or increasing the fracture stress should result in worse ductility. The scenario in the hard orientation NiAl single crystals is somewhat different as the mobility of <01 1> dislocations and the climb of <001> dislocations appear to determine the tensile ductility. However, to increase the high temperature strength of hard-oriented crystals, the mobility of <01 1> dislocations and the climb of <001> dislocations have to be impeded and this should result in a higher BDTT. Besides, the number of independent slip systems available with <01 1> slip is even less than that with <001> slip and, therefore, the activation of <01 1> slip systems in the absence of others is not expected to provide good toughness levels at lower temperatures. In contrast, <111> slip can provide sufficient independent slip systems to accommodate any plastic deformation. However, <1 1 1> slip has not been observed in this study. The only other way that NiAl could have a sufficient number of independent slip systems is by activating <001> and <01 1> slip simultaneously. However, the large difference between the CRSS's of <01 1> and <001> slip prevents their

PAGE 140

133 simultaneous activation at RT and hence prevents the ductilization of NiAl at RT. Above -800K, the CRSS of <01 1> slip decreases to a level similar to that of <001> slip (Figure 4.10) and, therefore, the simultaneous activation of both <001> and <01 1> slip is possible; this may partially account for the ductility of polycrystalline NiAl at high temperatures if the mobility of <01 1> dislocations is not too low. Therefore, it seems impossible to improve the low temperature ductility and high temperature strength of NiAl crystals simultaneously and, therefore, it is unlikely that this material will ever be used as a high temperature structural material. (010) (hkO) Figure 4.45 Schematic illustration showing the traces of dislocation bands, traces of debris and elongated loops when viewed along [001] zone for tensile axis are along [557] [hkl] and [01 1] The traces of primary slip planes responsible for these features are marked accordingly.

PAGE 141

134 40 r Temperature (K) Figure 4.46 Deviation angle vs. testing temperature (data compilation from (25)). The deviation angle is the angle between the observed slip plane determined by surface slip traces and the {110} slip plane. Data obtained from <123> tensile testing. It can be seen that the data can be extrapolated to OK where no cross slip is possible so slip will be confined to the {110} plane. The deviation of high temperature data from the line may be due to the increased effects of dislocation climb. T(K) or£" 1 Figure 4.47 Illustration of tensile ductility as a function of temperature or strain rate.

PAGE 142

135 Figure 4.48 Schematic illustration of forming new dislocation source via double cross slip, h > h. = 0.25 —

PAGE 143

CHAPTER 5 SUMMARY AND CONCLUSIONS The mechanical properties and slip behavior of high purity single crystalline NiAl over a range of temperatures (77K-1073K) have been examined. This study was aimed at revealing the details of the slip behavior, examining the effects of heat treatment and composition on this behavior, and clarifying the controversial issues concerning the reported slip behavior in both soft and hard oriented NiAl single crystals. It was also the purpose of this study to provide a more general understanding of the deformation mechanisms, especially the tensile ductility, of single crystalline NiAl by correlating the detailed slip behavior with the corresponding mechanical properties. Based on the results of this study, the following conclusions were drawn: 1. It was confirmed again by the results of this study that NiAl single crystals deform by either <00 1 > { 1 1 0 } or <001>{010} systems at all temperatures and obey Schmid's Law when the crystal is oriented along various soft orientations. Furthermore, it was found that in high purity NiAl, the <001>{ 1 10} and <001>{010} slip systems have the similar CRSS. It was also confirmed that the cross slip of <001> dislocations is easy and profuse and it was found that, due to the easy glide and easy cross slip of <001> dislocations, the slip is localized in the form of dislocation bands. Due to these characteristics, it is believed that the ductility of soft-oriented, high purity NiAl single crystals is determined by the rate of increase in the local stress concentrations which in turn, depends on the competition between the recovery mechanisms (cross slip and/or climb) and the local work hardening. Higher temperatures and slower strain rates, which enhance the effects of cross slip and climb, lower initial yield stress (higher purity) and lower strain 136

PAGE 144

137 hardening rates leads to higher elongation prior to reaching the cleavage stress. Specifically, it was found that: (a) The CRSS for the <001>{ 1 10} and <001>{010} systems is between 50 and 70 MPa, which is considerably lower than that of commercial purity NiAl. It was also noted that, for specimens cut from the same single crystal, there is typically less than 5% difference in the measured CRSS for these two slip systems. The work hardening rate of crystals oriented along single slip directions is typically very low and that of crystals oriented along [01 1] is approximately four times higher. (b) When oriented along [557] where only a single [001](Tl0) slip system operates prior to fracture, the high purity NiAl crystals exhibit a gradual increase in tensile elongation from 0% to -50% with increasing temperature from 77 to 473 K at a constant strain rate of 10" 4 s" 1 . At room temperature, over 30% elongation can be achieved using a strain rate of 10 4 s' 1 due to the lower CRSS for <001> slip and the fewer obstacles for dislocation glide in this material. (c) The <001> dislocations are easy to generate and glide at all the temperature studied. The dislocation substructures in the deformed samples indicate that, when gliding on { 1 10} planes, <001> edge segments appear to have lower mobility than <001> screw segments and the difference between the mobility of edge and screw <001> components appears to increase with decreasing temperature. It is also observed that the activity of <001> cross slip increases with temperature from little activity at 77K to large scale activity at RT. At and above room temperature, the slip was localized and occurred by the formation and broadening of dislocation bands. A high density of debris and prismatic loops elongated in [1 10] directions were also observed and mainly distributed in the dislocation bands. It is proposed that these loops are formed mostly via a double cross slip and jog dragging mechanism, and the broadening of dislocation bands via double cross slip.

PAGE 145

138 (d) The effect of heat treatment (cooling rate) on the slip behavior of [557] oriented specimens has also been investigated. When the [557] oriented specimens that were water quenched from 1273K were compared with the specimens furnace cooled from the same temperature, it was clear that the glide of <001> dislocations was considerably more difficult. The slip behavior of <001> dislocations was also more complicated in the water quenched specimens as secondary slip occurred and the traces of debris indicates slip on both {110} and {010} type planes, although the primary slip system remains as [001 ]( 1 10). The changes in the slip behavior of water quenched crystals are believed to be related to the dislocation-vacancy interactions. (e) Dislocation subboundaries were found in specimens deformed at temperatures as low as 673 K at a strain rate of 10 4 s "', indicating active dislocation climb at this temperature. All of the observed dislocation substructures in the specimens deformed at intermediate to high temperatures indicate that recovery had taken place and, accordingly, high tensile elongations were achieved at these temperatures when a balance between recovery and strain hardening was reached. Nevertheless, the dominating slip system is still [001](Tl0) and, therefore, the fracture remains brittle even after elongations of over 1 00% were achieved. (f) The dislocation substructures in both [T23] and [011] oriented specimens are found to have similar features as those in the [557] oriented specimens, i.e., slip localization (dislocation bands) and elongated prismatic loops along [1 10] for the [T23] specimens and [100] for the [011] specimens, indicating the similar slip behavior for all soft orientations. [T23] specimens were observed to deform by [001](1 10) slip, although local cross slip was abundant, especially at elevated temperatures. The slip systems in [01 l]-oriented crystals were confirmed to be [010](001) and [001](010) and dislocations of both types usually tangle together consistent with the higher work hardening rate of specimens along this orientation.

PAGE 146

139 2. The deformation of hard oriented NiAl single crystals at temperatures at or above 573K was confirmed by TEM observations to be by <01 1>{01 1 } slip. The BDTT of the high purity hard orientation NiAl was found to be between 523K and 573K at a strain rate of 10" 4 s"'. Immediately above the BDTT, the tensile elongation increases from less than 2% to over 20%, although the fracture remains brittle to high temperatures. It is proposed that the activation of <01 1>{01 1 } slip systems and the recovery of dislocation substructures enhanced by the climb of <001> dislocations result in the sudden increase in tensile ductility. Specifically, the following conclusions are drawn based on the results of this study. (a) NiAl and NiAl-0.3at%Si single crystals deform by <0 1 1 > { 0 1 1 } slip at or above 573K. (b) <011> dislocations have a much lower mobility (higher CRSS) than <001> dislocations and exhibit a skew configuration with a strong preference for the directions to lie along <1 1 1> in their slip planes. This kind of skew configuration can be explained with the kink motion of <1 1 1> segments. Therefore, it is proposed that the sudden increase in the tensile ductility of hard-oriented NiAl single crystals is due to the thermally activated kink motion which enhances the mobility of <01 1> dislocations. (c) Si does not strongly affect the slip behavior and slip transition of <001> oriented NiAl crystals, although the addition of Si appears to lead to an increase in the tensile elongation at temperatures above 573K. (d) The "BDTT" of <001> oriented NiAl single crystals is due to the activation of <011> slip systems and not due to the onset of the bulk climb of <001> dislocations. However, the climb of <001> dislocations is believed to enhance the tensile ductility by offering faster dynamic recovery of the dislocation substructures.

PAGE 147

LIST OF REFERENCES 1. R. Darolia, /. Metals 43,44(1991). 2. A. Ball and R. E. Smallman, Acta Metall. 14, 1517 (1966). 3. R. Bowman, R. Noebe and R. Darolia, HITEMP Review , NASA conference Pub. 10039,47-1, (1989). 4. R. D. Field, D. F. Lahrman and R. Darolia, Acta Metall. Mater. 39, 2961 (1991). 5. R. J. Wasilewski, Trans. Metall. Soc. AIME 236, 455 (1966). 6. T. Takasugi, S. Watanabe and S. Hanada, Mater. Sci. Eng. A 149, 183 (1992). 7. R. T. Pascoe and C. W. A. Newey, Phys. Stat. Sol. 29, 357 (1968). 8. R. J. Wasilewski, S. R. Butler and J. E. Hanlon, Trans. Metall. Soc. AIME 239, 1357 (1967). 9. R. T. Pascoe and C. W. A. Newey, Metal Sci. J. 2, 138 (1968). 10. E. M. Grala, Mechanical Properties of Intermetallic Compounds , J. H. Westbrook, Ed., Wiley, New York , p. 359 (1960). 11. D. F. Lahrman, R. D. Field and R. Darolia, High Temperature Ordered Intermetallic Alloys IV , MRS Symp. Proc. Vol. 213, 603 (1991). 12. R. D. Noebe, R. R. Bowman and M. V. Nathal, Inter. Mater. Rev. 38, 193 (1993). 13. J. D. Cotton, R. D. Noebe and M. J. Kaufman, Intermetallics I, 3 (1993). 14. D. R. Johnson, S. M. Joslin, B. F. Oliver, R. D. Noebe and J. D. Whittenberger, First International Conference on Processing Materials for Properties , edited by H. Henein and T.Oki, TMS, Warrendale, PA , 865 (1993). 15. V. L Levit, I. A. Bui, J. Hu and M. J. Kaufman, Scripta Mater. 34, 1925 (1996). 16. R. Darolia, D. F. Lahrman and R. Field, Scripta Metall. Mater. 26, 1007 (1992). 17. M . Weaver, Investigation of Strain Aging in The Ordered Intermetallic Compound Beta-NiAl , Ph.D. Dissertation, University of Florida (1995). 140

PAGE 148

141 18. M. Weaver, R. D. Noebe, J. J. Lewandowski, B. F. Oliver and M. J. Kaufman, Mater. Sci. Eng. A 192/193, 179 (1996). 19. M. Weaver, R. D. Noebe, J. J. Lewandowski, B. F. Oliver and M. J. Kaufman, Intermetallics 4, 533 (1996). 20. M. Weaver, R. D. Noebe and M. J. Kaufman, Scripta Mater. 34, 941 (1996). 21. J. E. Hack, J. M. Brzeski and R. Darolia, Mater. Sci. Eng. A 192/193, 268 (1995). 22. M. H. Loretto and R. J. Wasilewski, Phil. Mag. 23, 1311 (1971). 23. W. R. Kanne Jr, P. R. Strutt and R. A. Dodd, Trans Metall. Soc. AIME 245, 1259 (1969). 24. A Ball and R.E. Smallman, Acta Metall. 14, 1 349 (1966). 25. T. Takasugi, J. Kishino and S. Hanada, Acta Metall. Mater. 41, 1021-103 1 (1993). 26. F. Ebrahimi, A. Gomez and T. G. Hicks, Scripta Mater. 34, 337-342 (1996). 27. J. T. Kim, On the Slip Behavior and Surface Film Effects in B2 Ordered NiAl Single Crystals , Ph.D. Dissertation, The University of Michigan (1991). 28. R. D. Field, D. F. Lahrman and R. Darolia, Acta Metall. Mater. 39, 2951 (1991). 29. Y. Q. Sun, G. Taylor, R. Darolia and P. M. Hazzledine, High-Temperature Ordered Intermetallic Alloys VI , MRS Symp. Proc. Vol. 364, 261 (1995). 30. K. R. Forbes, Mechanisms of High Temperature Deformation of NiAl single Crystals , Ph.D. dissertation, University of Stanford (1994). 31. W. S. Walston and R. Darolia, High Temperature Ordered Intermetallic Alloys V , MRS Symp. Proc. Vol. 288, 237 (1993). 32. P. Nash and D. R. F. West, Metal Science 15, 347 (1981). 33. P. Nash, M. F. Singleton and J. L. Murray, Phase Diagrams of Binary Nickel Alloys , P. Nash, Ed., Vol. 1, (1991). 34. H.-D. Dannohl and H. L. Lukas, Z. Metallic. 65, 642(1974). 35. E.-T. Henig and H. L. Lukas, Z. Metallk. 66, 98 (1975). 36. K. Sumiyama, Y. Hirose and Y. Nakamura, Phys. Stat. Sol. (A) 114, 693 (1989). 37. A. G. Fox and M. A. Tabbernor, Acta Metall. Mater. 39, 669 (1991). 38. N. Rusovic and H. Warlimont, Phys. Stat. Sol. (a) 44, 609 (1977).

PAGE 149

142 39. A. Nagasawa and Y. Ueda, /. Phys. Soc. Japan 45, 1249 (1978). 40. K. Enami, J. Hasunuma, A. Nagasawa and S. Nenno, Scripta Metall. 10 879 (1976). 41. S. B. Maslenkov, A. I. Kozlenkov, S. A. Filin and A. I. Shulgin, Phys. Stat Sol (b) 123, 605 (1984). 42. M. J. Cooper, Phil. Mag. 89, 805(1963). 43. A. Taylor and N. J. doyle, /. Appl. Crystallogr. 5, 201 (1972). 44. P. Georgopoulos and J. B. Cohen, Acta Metall. 29, 1535 (1981). 45. R. R. Bowman, R. D. Noebe, R. D. Raj and I. E. Locci, Metall. Trans 23A 1493 (1992). 46. D. B. Miracle, Acta Metall. Mater. 41, 649 (1993). 47. H. L. Fraser, M. H. Loretto and R. E. Smallman, Phil. Mag. A 28, 667 (1973). 48. I. Baker, Mater. Sci. Eng. A192/193, 1 (1995). 49. I. Baker and P. R. Munroe, High Temperature Aluminides and Intermetallics . edited by S.W.Whang, C.T.Liu, D.P.Pope and J.O.Stiegler, TMS, 425 (1990). 50. W. A. Rachinger and A. H. Cottrell, Acta Metall. 4, 1517(1956). 51. D. I. Potter, Mater. Sci. Eng. 5, 201 (1969/70). 52. A. H.E. Foreman, Acta Metall. 3, 322(1955). 53. C. H. Lloyd and M. H. Loretto, Phys. Stat. Sol. 39, 163 (1970). 54. D. B. Miracle, Acta Metall. Mater. 39, 1457(1991). 55. U. Glatzel, K. R. Forbes and W. D. Nix, Phil. Mag. A 67, 307 (1992). 56. J. D. Eshelby, Phil. Mag. 40, 903 (1949). 57. M. Yamaguchi and Y. Umakoshi, Scripta Metall. 9, 637 (1975). 58. D. Farkas, R. Pasianot, E. J. Savino and D. B. Miracle, High Temperatur e Ordered Intermetallic Alloys IV MRS Symp. Proa, Vol. 213, 223 (1991). 59. T. P. Parthasarathy, D. M. Dimiduk and G. Saada, High Temperature Ord ered Intermetallic Alloys V MRS Symp. Proc. , Vol. 288, 311 (1993). 60. M. J. Mills and D. B. Miracle, Acta Metall. Mater. 41, 85 (1992).

PAGE 150

143 61. M. A. Crimp, S. C. Tonn and Y. Zhang, Mater. Sci. Eng. A 170, 95 (1993). 62. P. Veyssiere and R. Noebe, Phil. Mag. A 65, 1 (1992). 63. M. J. Mills and D. B. Miracle, Acta Metall. Mater. 41, 85 (1993). 64. R. D. Field, D. F. Lahrman and R. Darolia, High Temperature Ordered Intermetallic Alloys IV , MRS Svmp. Proc. Vol.213, 255 (1991). 65. H. L. Fraser, , R. E. Smallman and M. H. Loretto, Phil. Mag. A, 28, 651 (1973). 66. R. D. Noebe, The Effect of Various Metallurgical Parameters on the Flow and Fracture Behavior of Polycrystalline NiAl Near the Brittle-To-Ductile Transition NASA Technical Memorandum 106534, (1994). 67. J. D. Cotton, R. D. Noebe and M. J. Kaufman, Intermetallics 1, 1 17-126 (1993). 68. J. Hu, V. Levit and M. J. Kaufman, High Temperature Ordered Intermetallic Alloys VI, MRS Symp. Proc, Vol. 364, 297 (1995). 69. W. Yang, R. A. Dodd and P. R. Strutt, Metall. Trans. 3, 2049 (1972). 70. J. E. Epperson, K. W. Gerstenberg and D. Berner, Phil. Mag. A 38, 529 (1978). 71. A. Parthasarathi and H. L. Fraser, Phil. Mag. A 50, 89 (1984). 72. A. Ball and R. E. Smallman, Acta Metall. 16, 233 (1968). 73. J. M. Koch and C. Koenig, Phil. Mag. B 54, 177 (1986). 74. H. L. Fraser, M. H. Loretto, R. E. Smallman and R. J. Wasilewski, Phil Mas A 28, 639 (1973). 75. W. J. Yang and R. A. Dodd, Scripta Metall. 8, 237 (1974). 76. J. E. Eibner, H.-J. Engell, H. Schultz, H. Jacobi and G. schlatte, Phil. Mas. 31 739(1975). 77. P. Nagpal and I. Baker, Metall. Trans. 21A, 2281 (1990). 78. J. Winston, The Effect of Orientation, Temperature, And Strain Rate on The Mechanical Properties of NiAl Single Crystals . M.S. thesis, University of Florida (1995). J 79. R. Z. Von Mises, Angew. Math. Mech. 8, 161 (1928). 80. K. H. Hahn, K. Vedula, Scripta Metall. 23, 7 (1989). 81. E. M. Schulson, High Temperature Ordered Intermetalli c Alloys, MRS Svmp Proc, Vol. 39, 193 (1985). y F '

PAGE 151

144 82. E. M. Schulson, Res. Mech. Lett. 1, 1 1 1 (1981). 83. R. D. Field, D. F. Lahrman and R. Darolia, High Temperature Ordered Intermetallic Alloys V , MRS Symp. Proc, Vol. 288, 423 (1993). 84. J. M. Brzeski, J. E. Hack, R. Darolia and R. D. Field, Mater. Sci. Eng. A170 11 (1993). 85. T. Takasugi, S. Wantanabe and S. Hanada, Acta Metall. Mater. 41, 1009-1020 (1993). 86. G. W. Grove and A. Kelly, Phil. Mag. 19,977-986(1969). 87. R. D. Noebe and A. Garg, Deformation Behavior of the Single Crystal NiAl Alloy D176 , NASA Technical Memorandum, (1995). 88. J. T. Kim, R. Gibala, High Temperature Ordered Intermetallic Alloys IV , MRS Symp. Proc, Vol. 213, 261 (1991). 89. M. J. Mills, M.S. Daw, J. E. Angelo and D. B. Miracle, High Temperature Ordered Intermetallic Alloys VI MRS Vol. 364 , 401 (1995). 90. F. Ebrahimi and S. Shrivastava, Acta Metall. Mater, in press (1997). 91. V.I. Levit, J. Hu, L A. Bui, J. S. Winton and M. J. Kaufman, "Challenges in the Development and Application of Beta-NiAl as an Engineering Material", Paper presented at the Engineering Foundation Meeting, Davos, Switzerland, May 19-24 (1996). 92. S. Shrivastava, Brittle-To-Ductile Transition In NiAl Single Crystals , Ph.D. dissertation, University of Florida (1997). 93. J. R. Low and A. M. Turkalo, Acta Metall . 10, 215 (1962). 94. R. L. Segall, P. G. Partridge and P. B. Hirsch, Phil. Mag. 6, 1493 (1961). 95. J. T. Fourie and H. G. F. Wilsdorf, /. Appl. Phys. 31, 2219 (1960). 96. P. B. Price, Phil. Mag. 5, 873 (1960). 97. J. Washburn et al., Phil. Mag. 5, 991 (1960). 98. W. C. Dash, /. Appl. Phys. 29, 705 (1958). 99. W. G. Johnston and J. J. Gilman, /. Appl. Phys. 31, 632 (1960). 100. A. S. Tetelman, Acta Metall. 10, 813(1962). 101. S. A. Maloy, G. T. Gray III and R. Darolia, Mater. Sci. Eng. A 192/193 249 (1995).

PAGE 152

145 102. D. Hull, Introduction to Dislocations , 2nd ed. Robert Maxwell. M.C.. p 161 (1975). 103. G. E. Dieter, Mechanical Metallurgy , 3rd ed. McGraw-Hill Book Company, pi 28 (1986). 104. V. I. Levit, J. Hu and M. J. Kaufman, Structural Intermetallics 1997 , edited by M.V. Nathal, R.Darolia, C.T.Liu, P.L.Martin, D.B. Miracle, R. Wagner, and M. Yamaguchi, TMS, 683 (1997). 105. L. K. France, C. S. Hartley and C. N. Reid, Metal Sci. J. 1, 65 (1967). 106. C. Vailhe and D. Farkas, High Temperature Ordered Intermetallic Alloys VI , MRS Symp. Proc. Vol. 364, 395 (1995). 107. J. R. Low and R. W. Guard, Acta Metall. 7, 171 (1959). 108. H. Neuhauser, Dislocation In Solids , Vol. 6, North-Hollan Publishing Company, p321 (1983). 109. P. D. K. Nathanson, P. J. Jackson and D. R. Spalding, Acta Metall. 28, 823 (1980). 110. A. Seeger and H. Trauble, Z Metallk. 51, 435 (1 960). 111. H. Wiedersich, /. Appl. Phys. 33, 854 (1962). 112. V. I. Levit, J. S. Winton, Y. N. Gornostyrev and M. J. Kaufman, Recrystallization-96 , Proceedings of the Third International Conference on Recrystallization and Related Topics, edited by T.R. McNelley, Monterey, CA, 637-644 (1996). 113. V. V. Tokii, Phys. Stat. Sol. 44, 767 (1977). 1 14. J. P. Hirth and J. Lothe, Theory of Dislocations , Krieger Publishing Company, (1982). 115. R. W. Balluffi, Phys. Stat. Sol. 42, 1 1 (1970). 116. C. St John, Phil. Mag. 32, 1193 (1975). 117. R. T. Pascoe and C. W. A. Newey, Met. Sci. J. 5, 50 (1971).

PAGE 153

BIOGRAPHICAL SKETCH Jian Hu was born on November 2, 1962, in Shanghai, China. Upon completing high school, he attended Shanghai Jiao Tong University, where he received a B.S. degree in materials science and engineering in 1983. He then worked in industry for one year. In September 1984 he enrolled at the South China University of Technology and received his M.S. degree in Metallurgy in July, 1987. From 1987 to 1992, he worked for the Shanghai Iron and Steel Research Institute as a research engineer. He enrolled at the University of Florida in January, 1993 and completed his Ph.D. dissertation study in December, 1997. 146

PAGE 154

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Michael J. Kaufman, [upairman Professor of Materials Science and Engineering I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. l nhert Reed-Hill / Robert Reed-Hill Professor of Materials Science and Engineering I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Robert T. DeHof Professor of Materials Science and Engineering I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Fereshteh Ebrahimi Associate Professor of Materials Science and Engineering I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. — Ashok V. Kumar ~~ Assistant Professor of Mechanical Engineering

PAGE 155

This dissertation was submitted to the Graduate Faculty of the College of Engineering and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1997 J— Winfrc Winfred M. Phillips Dean, College of Engineering Karen A. Holbrook Dean, Graduate School