The influence of chromium on structure and mechanical properties of B2 nickel aluminide alloys

MISSING IMAGE

Material Information

Title:
The influence of chromium on structure and mechanical properties of B2 nickel aluminide alloys
Physical Description:
vi, 151 leaves : ; 29 cm.
Language:
English
Creator:
Cotton, James Dean, 1961-
Publication Date:

Subjects

Subjects / Keywords:
Chrom   ( swd )
Mechanische Eigenschaft   ( swd )
Nickelaluminide   ( swd )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 140-150).
Statement of Responsibility:
by James Dean Cotton.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001709815
notis - AJC2100
oclc - 25541190
Classification:
rvk - ZM 3200
System ID:
AA00003293:00001

Full Text











THE INFLUENCE OF CHROMIUM ON STRUCTURE AND
MECHANICAL PROPERTIES OF B2 NICKEL ALUMINIDE ALLOYS














BY


JAMES DEAN COTTON


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



UNIVERSITY OF FLORIDA


1991














ACKNOWLEDGEMENTS


Few individuals achieve the distinction of a Doctor of Philosophy degree without

the emotional, technical, and financial support of others; I am happy to recognize my own.
First, I thank my parents for their unwavering support of my goals. In addition, I have a

number of friends in Seattle who encouraged me to attempt this degree, in spite of my own

doubts. I particularly remember the encouragement of Steve Spear, Melanie Zerbe, and

Dave House. The companionship of the following people at the University of Washington

and the University of Florida will also be fondly recalled: Dick Kennish (those "required"

weekly ski trips), Kenji Kojima and the UW Intramural Volleyball Champions of 1988-

89, and Andy Duncan and Mark Weaver (who can quit drinking coffee anytime they want

to). The following NASA LRC employees were invaluable sources of of technical and/or

recreational advice: Randy Bowman, Ivan Locci, Mike Nathal, Bob Miner and Hugh

Gray. Thanks also go to Mark Behbehani for his excellent assistance in mechanical testing.

I especially acknowledge the technical guidance, suggestions and friendship of Ron Noebe.

Of course, the financial support of the University of Washington, the University of Florida,

DARPA and NASA are graciously acknowledged.

I have the utmost respect and gratitude for Professor Mike Kaufman, who always
had the insight to ask of me more than I felt capable of, and thereby helped me to grow

immensely. His passion for understanding (and ability to locate funding) has been an

inspiration.

Finally, the love and support of my wife, Maria, are recognized. The patience and
tolerance she has demonstrated, the counsel she has offered, and the good and bad times

we have shared during this period of our life, have been an important part of this goal.















TABLE OF CONTENTS



ACKNOWLEDGMENTS........................ ....................................................... ii

ABSTRACT ................................................................................................................ v

CHAPTERS

1 INTRODUCTION................................................................................. 1

Background............................................................................................... 1
Approach................................................................................................... ...... 3

2 REVIEW OF THE LITERATURE............................................................ 4

Nickel-Aluminum Phase Equilibria....................................... ................ 4
Structure of NiAl......................................................................................... 5
Mechanical Properties of NiAl.................................................................. 6
Deformation and Fracture........................................................................... 16
Ternary Alloying Effects on Plastic Deformation...................................... 25

3 EXPERIMENTAL PROCEDURE....................................... ............... 37

Materials.................................................................................................... 37
Optical Microscopy................................................................................... 39
Mechanical Property Testing................................................................... 41
Transmission Electron Microscopy......................................................... 44
X-ray Diffraction Analysis...................................................................... 47

4 RESULTS.................................................................................................. 48

Alloy Compositions............................................... ................................. 48
Optical Microscopy................................................................................... 48
Mechanical Properties...................................................................................... 51
Transmission Electron Microscopy................................... .................. 70

5 DISCUSSION.............. .......... .............. ................................. .............. 105

Microstructural Evolution................................................................................ 105
Dislocation Configuration and Slip.................................................................... 108
Composition-Structure-Property Relationships.......................................... 113
Site Preference of Chromium.......................................................................... 132



iii









6 SUMMARY AND CONCLUSIONS................................................................ 134

APPENDICES

A TRUE STRESS-TRUE STRAIN CURVES.................................................... 136

B RAW BURGERS VECTOR ANALYSIS DATA...................................... 138

REFERENCES.................................................................................................... 140

BIOGRAPHICAL SKETCH...................................................................................... 151















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


THE INFLUENCE OF CHROMIUM ON STRUCTURE AND
MECHANICAL PROPERTIES OF B2 NICKEL ALUMINIDE ALLOYS

By

James Dean Cotton

December, 1991



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


Major obstacles to the use of NiAl-base alloys and composites are low ductility and

toughness. These shortcomings result, in part at least, from a lack of sufficient slip

systems to accommodate plastic deformation of polycrystalline material (the von Mises

Criterion). It has been reported that minor additions of chromium to polycrystalline NiAl

cause the predominant slip system to shift from the usual <001>{110} to <111>{112}. If

this is true, then a major step toward increasing ductility in this compound may be realized.

The purpose of the present study was to verify this phenomenon, characterize it with

respect to chromium level and nickel-to-aluminum ratio, and correlate any change in slip

system with microstructure and mechanical properties. Compression and tensile

specimens were prepared from alloys containing 0 to 5 at.% chromium and 45 to 55 at. %

aluminum. Following about one percent strain, TEM foils were produced and the slip

systems determined using the g-b-0 invisibility criterion. Contrary to the previous results,

chromium was found to have no effect on the preferred slip system in of any of the alloys










studied. Possible reasons for the inconsistency of the current results with previous work

are considered. Composition-structure-property relationships are discerned for the alloys,

and good correlations are demonstrated in terms of conventional strengthening models for

metallic systems.













CHAPTER 1
INTRODUCTION


Background

A class of materials termed "intermetallics" has been the subject of an increasing

amount of study for the past decade. There are several reasons for this interest, most of

which revolve around a need for stable high-temperature ("high" with respect to most

metals) materials. The intermetallics are considered to occupy a position somewhere

between metals and ceramics, due to a significant covalent contribution to the overall

bonding of the compound which lends thermal stability. This often causes an intermetallic

to have a melting point above one, or both, of its pure constituent elements. In addition,

intermetallics often exhibit, relative to most pure metals, higher room temperature strength,

higher moduli and lower thermal expansion coefficients. Thermal and electrical

conductivities can exceed competing metallic alloys, due to the increased rigidity of the

structure. However, the same covalent component which creates such attractive properties

will, by nature, also reduce the ability of the material to undergo plastic deformation by

dislocation motion. Thus, the ductility and toughness of intermetallic compounds are

usually low compared with metals. This latter deficiency has been a major obstacle in the

development of intermetallics and is the reason for the current study.

One very important application for intermetallics is high-temperature components

of advanced turbine engines, such as blades and vanes [1, 2]. For such applications, B2

nickel aluminide (NiAl) is a material of choice and the next logical step in materials

evolution beyond current superalloys. Nickel aluminide has a high melting point (300K

above superalloys), a low density (75 percent that of superalloys or Ni3AI), a high thermal









conductivity (as much as eight times that of superalloys), a wide solubility range, good

processability, and is inexpensive [3].

In spite of these advantages, NiAl is brittle and rarely attains a tensile ductility in

excess of 2 percent at room temperature [4, 5, 6, 7, 8]. Room-temperature plastic

deformation in NiAl occurs predominantly by dislocation glide on the <100>(011}

systems [9, 10, 11], of which only three are independent [12]. According to the von Mises

Criterion [13], a minimum of five independent deformation mechanisms are required for

arbitrary deformation of any three-dimensional body. This requirement restricts plastic

flow in polycrystalline NiAl, as each grain must undergo arbitrary deformation to maintain

shape compatibility with the adjacent grains, and is at least one reason why it is not ductile.

However, if other slip systems could be activated, such as those which operate in body-

centered-cubic metals (<111>{0OT1 ), then sufficient slip systems would be available and a

major obstacle to ductility and toughness will have been removed.

One avenue to activating other slip systems that has been suggested is through

ternary alloying [14, 15, 16]. In fact, Law and Blackburn [15] reported in 1985 that the

addition of five atom percent chromium to NiAl caused the primary Burgers vector to

change from <100> to <111>. However, no increase in tensile ductility was observed.

This result is intriguing and several questions are obvious. If the Burgers vector was truly

altered and sufficient slip systems were available, then what other factor is restricting

plastic flow? Are there other compositions in this ternary system which may demonstrate

similar behavior? The primary question, fundamental to any scientific study is, "Are the

results reproducible?" As of the time of this writing, there have been no other studies

which confirm the effect of chromium on plastic flow in polycrystalline NiAl. Thus, the

objective of the current study is to examine, in a systematic fashion, the effect of chromium

on slip and mechanical properties in NiAl.












Since Law and Blackburn [15] initially studied cast alloys, and the room

temperature solubility of chromium in NiAl is less than two atom percent [17, 18], the

chromium content of the alloys in the current study was varied from zero to five atom

percent (0, 1, 2, 5). To examine the effects of stoichiometry, the aluminum content was

also varied, from 45 to 55 percent. A set of 15 alloys within this composition range was

cast and homogenized, including one exactly at the composition reported to exhibit <111>

slip [15]. To study processing effects and improve specimen integrity, five of the

compositions (binary stoichiometric NiAl and four aluminum-lean ternaries) were also

produced in extruded form. Both the cast and the extruded alloys were utilized for an

overall survey of Burgers vectors in this system. In addition, the relationships between

composition, processing, dislocation type and mechanical properties were examined. The

greater microstructural integrity of the extrusions allowed the determination of tensile

properties as well as brittle-to-ductile-transition-temperatures.














CHAPTER 2
REVIEW OF THE LITERATURE


This review covers topics germane to understanding the chemical and structural

factors controlling plastic deformation in NiAl under ambient conditions. The primary

topics are phase equilibria, mechanical behavior, deformation and fracture, and alloying

effects. The latter topic deals with the effect of ternary additions on slip behavior and

mechanical properties and is mainly confined to single phase alloys of the B2 structure.


Nickel-Aluminum Phase Equilibria

The binary nickel-aluminum phase diagram is shown in Figure 2-1. Five

intermetallic phases are present in this system. Of the five, the compound NiAl exhibits

the highest melting point, 1911K. Nickel aluminide also displays an unusually wide range

of solubility, an indication of high stability, spanning more than 20 atomic percent (all

compositions will be given in atomic percent) at 14000C. Deviations from stoichiometry

are accommodated by the substitution of nickel atoms onto aluminum sites in nickel-rich

compositions (antisite defects) and by the formation of vacancies on the nickel sites in

aluminum-rich compositions [19]. The inability of aluminum atoms to occupy the nickel

sites is usually explained by the larger atomic size of aluminum atoms. Vacancies are

formed because the overall alloy composition can only be maintained by the elimination of

a proportionate number of nickel atoms. These antisite and vacancy point defects have

been shown to locally order such that no defect has a similar defect as a first neighbor [20]

which leads to clusters of Ni2Al and Ni2A13, respectively [21].










Structure of NiAl

Beta NiAl crystallizes in the CsCI (B2) structure, in which one atom type resides at

the center position of a cubic unit cell and the other at the covers, thus comprising an

ordered body-centered cubic (BCC) lattice [23]. The structure is shown in Figure 2-2.

1800 i

1600


1400

1200

1000

800

600

300


10 20 30 40 50 60
Atomic Percent Ni


70 80 90 100
Ni


Figure 2-1. Binary Phase Diagram for Nickel-Aluminum System [22]





Ni 0


Figure 2-2. NiAl B2 Structure









This compound is a congruently melting Hume-Rothery phase with an electron-to-

atom ratio of 1.5, similar to CuZn, and is strongly ordered [24]. In addition, there is

evidence that short-range order exists in the melt [25]. This strong tendency to order is

indicative of a high ordering energy and a preference for bonding between aluminum and

nickel atoms. X-ray and electron structure factor measurements [26, 27, 28, 29] provide

evidence of a high degree of bond directionality, indicating electron concentration in <111>

directions between first-neighbor aluminum-nickel pairs and a mixture of both metallic and

covalent bonding. This large attraction between aluminum and nickel atoms accounts for

many of the unusual properties of NiAl.


Mechanical Properties of NiAl

Polycrystalline


Strength and Ductility. Room-temperature mechanical properties of NiAl in

tension and compression are summarized in Tables 2-1 and 2-2, respectively.

The earliest report on the mechanical properties of NiAl is that of Wachtell [30],

which gives a room-temperature modulus-of-rupture for powder-processed material of

735 MPa, with nil ductility. This value is high compared to more recent reports and is

likely a result of impurities introduced by powder processing. Later work by Maxwell and

Grala [6, 31, 32] reported a much lower fracture strength for cast material, about 100 MPa,

and zero ductility also. The first report of ambient temperature ductility in NiAl was by

Rozner and Wasilewski [33], who measured about 4 percent elongation and a yield

strength of 179 MPa in cast and extruded material. Quite recent studies have essentially

confirmed the ambient ductility as being about 1-2 percent, while the yield strength varies,

depending on the actual study, from 115 to 235 MPa, with a mean value of 167 MPa [4, 7,

33, 34, 35, 36, 37]. The differences in strength may result from variations in interstitial

content, such as carbon (a common impurity in nickel) [35] or stoichiometry. The grain









size has been shown experimentally to have little effect on either the yield or fracture

strengths of stoichiometric NiAl at ambient temperatures [36, 37], although fracture

mechanical models predict otherwise [34, 38, 39, 40, 41]. The possible effect of grain size

on mechanical properties is discussed in further depth below.


Table 2-1. Room-Temperature Tensile Properties of Polycrystalline NiAl

TYS UTS Max. Grain Size Processing
Reference Year (MPa) (MPa) Strain (%) (pm) Al (At.%) Method
[30] 1952 --- 738 0 -40 -50 PM, HP
[32] 1957 --- 103 0 -200 -50 cast
[33] 1966 179 207 4 no data 50.5 cast, extr.
[4] 1989 235 324 2.5 10-16 50.3 cast, extr.
[49] 1989 --- 283 0 33 50.6 PM, extr.
[35] 1990 154 228 2 30 -50 cast, extr.
[36] 1990 115 220 2.3 13 -50 cast, extr.
[34] 1991 220 400 0-1 5 -50 PM, extr.
[68] 1991 --- 350 0 10 50.6 PM, extr.


Table 2-2. Room-Temperature Compressive Properties of Polycrystalline NiAl


Note the correlation in Table 2-1 between ductility and processing by extrusion.

This correlation has been proposed to be due to a <111> fiber texture brought about by the

extrusion process and the retention of a mobile dislocation substructure [4]. As with many

metals, grains reorient during extrusion such that softer directions are aligned with the

extrusion axis; a portion of this texture is retained following recrystallization. When test

specimens are machined from the extrusions, the stress axis coincides with the extrusion

axis and thus the soft orientation. This will tend to allow a lower yield stress and, for a

given fracture stress, slightly higher ductility. In addition, higher ductility is apparent for








cast and extruded material compared with consolidated powder, which may be a result of

contamination acquired during powder processing. Any contaminant phases may operate

as critical flaws in the consolidated product.

Hardness. The room-temperature hardness of stoichiometric and near-
stoichiometric NiAl is in the range 250 320 kgmm-2 (Vickers hardness), depending upon

the exact composition and impurity content [6, 23, 42, 43, 44, 45, 46, 47].

Effect of Temperature. A substantial amount of data exists for the mechanical

properties of NiAl as a function of temperature [6, 7, 23, 32, 33, 34, 48, 49], the most

extensive of which are that of Rozner and Wasilewski (tension) and Pascoe and Newey

(compression) [7, 33].

Stoichiometric NiAl displays a smooth decrease in yield strength with increasing

test temperature and shows no indication of the strength anomaly sometimes observed in

B2 compounds [50], Figure 2-3. The figure is adapted from Vedula and Khadikar [51]

and represents the combined data of several investigators [4, 33, 34, 49], which agree quite

well. The yield strength of polycrystalline NiAl is low, about 170 MPa (25 ksi) at room

temperature, which makes it comparable to moderately strong aluminum alloys.

Similar to BCC metals, NiAI exhibits a transition from brittle to ductile behavior

with increasing temperature, as illustrated in Figure 2-4 for several different studies. A

Brittle-to-Ductile-Transition-Temperature (BDTT) may be determined for polycrystalline,

stoichiometric NiAl to be about 600K (0.3 Tm), since the results indicating higher BDTTs

may be due to nonstoichiometric material or impurity effects, both of which raise the

tensile strength further above the fracture stress. It is quite interesting to note that the

BDTT is lower for more recent studies, probably an effect of cleanliness (interstitial

elements) or better compositional control. Other extrinsic effects which may limit apparent

ductility (raise the BDTT) include defects in powder-processed material, surface roughness

and misaligned tensile specimens. In addition, increasing the strain rate (typically at 104/s)

has the effect of raising the BDTT about 50K for each order of magnitude [34, 52].








300


200 1


100 -


200


400 600 800 1000 1200 1400


Test Temperature (K)

Figure 2-3. Tensile Yield Strength of Polycrystalline NiAl as a Function of Temperature.
Adapted from Vedula and Khadikar [51]


0 200 400 600 800 1000


1200 1400


Test Temperature (K)

Figure 2-4. Ductility (Tension) of Polycrystalline NiAl as a Function of Temperature
[32,33,34,49].


IF









The collective ultimate tensile and fracture strength data of several investigators are

plotted below in Figures 2-5 and 2-6. The behavior in compression is roughly regular,

with approximately a 100 MPa increase in fracture strength for each 100K decrease in

temperature. In contrast, the tensile fracture strengths show considerable variability and

disagreement between studies below 1000K. This is probably due to differences in the

defect size and population, as fracture strength is inversely related to maximum defect size.

Although comparison between tensile and compressive fracture strengths is not strictly

valid, it can be immediately seen from these plots that the large discrepancy is more

pronounced at temperatures below 875K. This is likely due to the intervention of fracture

in tension, which occurs at a lower stress than plastic deformation when at lower

temperatures. In compression, although cracks form, they propagate in a stable fashion,

and deformation can continue to greater strains without complete failure [34].


300

250

200

U 150

100

B 50


0 200 400 600 800 1000 1200 1400

Test Temperature (K)

Figure 2-5. Ultimate Tensile Strength of Polycrystalline NiAl as a Function of
Temperature [6, 32, 34, 49]












1200

1000 Pascoe and Newey


800 A
AA A

600
0 -Ball and Smallman
400

200


0 200 400 600 800 1000 1200 1400

Test Temperature (K)

Figure 2-6. Fracture Strength in Compression of Polycrystalline NiAl as a Function of
Temperature [7, 48]


Effect of Stoichiometry. The stoichiometry of the NiAl compound has been

shown to have a marked effect on the mechanical properties. Exact stoichiometry, i.e.

Ni/Al = 1, is associated with a maximum in ductility and a minimum in strength and

hardness. Westbrook [23] first indicated this effect by microhardness testing of arc-melted

buttons, Figure 2-7, and associated this behavior with the defect structures proposed by

Bradley and Taylor [19]. Nickel-rich compositions are hardened relative to stoichiometric

compositions by the presence of nickel atoms on aluminum sublattice sites (antisite

defects) while aluminum-rich compositions are hardened to a greater extent by the

occurrence of vacancies on the nickel sublattice. This behavior also extends to yield

strength, which shows a minimum at the stoichiometric composition, Figure 2-8 [51].









800


600



400



200,


45 50 55 60 65

Atomic Percent Ni

Figure 2-7. Hardness of NiAl as a Function of Stoichiometry and Temperature [23]


700
573 K
0 Tensile and
Compression Data

500


S 400

vacancy
300 antisite defects
200 defects

200 -


100 I I I '
45 47 49 51 53 55

Atomic Percent Al

Figure 2-8. Effect of Stoichiometry on Yield Strength of Polycrystalline NiAl [51]


298


El



, ,


Cu u


1073K


K









There have been studies which appear to indicate that mechanical property maxima
and minima occur for slightly aluminum-rich compositions [33, 51]. However, the
difficulty in the exact determination of aluminum content (usually + 1.0 atom percent) and
the occurrence of physical property inflections at the exact stoichiometric composition [53,
54] indicate otherwise.
Effect of Grain Size. Based on early work by Cottrell [55], Schulson [39]
suggested there may be a critical grain size, dc, below which semi-brittle materials may
exhibit tensile elongation. He derived the following expression:



dc =
IYK kG



where KIc is the plane strain fracture toughness, Y is a geometrical parameter (order of 1),
and Ky and oo are the empirical Hall-Petch parameters. In this model, large grains are

considered to develop larger stress concentrations at grain boundaries than small grains,
due to the greater pileup of dislocations at the boundary. Assuming that the stresses to
propagate slip across a boundary and to nucleate a microcrack are similar, larger grains
should propagate cracks at a lower applied stress. Experimental verification of this concept
was later reported by the same author [40, 41] for cast and extruded Ni-49A1. This alloy
was extruded at a low temperature to retain a fine recrystallized grain structure, and
subsequent annealing treatments were imposed to attain a range of grain diameters, from 5
to 140 pm. Tensile tests at temperatures ranging from 293 to 873K indicated the existence
of a critical grain size, which was about 20 gpm at 673K, Figure 2-9. However, the room-
temperature ductility remained at about 2 percent, even at the finest grain size of 5 gm.

This would seem to indicate the critical grain size for Ni-49A1 at room temperature is
somewhat less than 5 pgm. This is supported by the work of Noebe et al. [34] who, using a









critical J-integral model developed by Chan [38], indicated that the critical grain size for
NiAl was about 1 uim, and that only 5 percent elongation can be expected for a grain size of

0.1 im. Further work by Nagpal and Baker [36] with cast and extruded Ni-50A1 found no

dependence of yield strength on grain size at 300K and, although finer grain sizes did

produce greater elongation, only 2.3 percent was achieved for the finest size studied (13
im). Quite recent results by Baker et al. [37] for Ni-45, -48 and -50A1 again found little

dependence of yield strength on grain size for Ni-50Al, but an increasing dependence with

deviation from stoichiometry. Since these tests were conducted in compression, no

ductility was reported.


50 100

Grain Diameter (microns)


150


Figure 2-9. Effect of Grain Size on Ductility of Ni-49A1 [40]


There are current efforts to produce NiAl with a grain size less than 5 microns,
utilizing special processing such as mechanical alloying plus extrusion to achieve a 1 p.m

grain size, and inert gas condensation to develop a 10 nm ("nanocrystalline") grain size [43,

56]. No tensile properties have been reported as yet for these materials.










Single Crystals

Yield Strength. The 0.2 percent offset yield strength for single crystal NiAl in
several orientations is shown in Table 2-3 with data from several studies [7, 48, 57, 58, 59,

60]. A marked anisotropy in the yield strength is apparent, in spite of the differences

between studies, and the typical strength of [100] (hard) oriented crystals is six times that

of [111] (soft) oriented crystals. Such a large orientation dependence is a direct result of

the preferred slip systems which are, in turn, a result of the directional bonding of NiAl.

Despite the known elastic anisotropy of B2 compounds [61], there is little relationship

between the elastic and plastic properties in this respect. For example, both NiAl and

CuZn are B2 compounds and have Zener anisotropy factors of 3.28 and 4.95, respectively.

However, because of the relatively larger metallic component of the bonding in CuZn, its
yield strength ratio Cl[l00]/[111] ~ 0.5, versus 6 for NiAl [61]. Thus, the character of the

bonding has a large effect in determining the properties of B2 compounds [62, 63, 64].


Table 2-3. Yield Strengths (MPa) of Single Crystal NiAl in Several Orientations at 300K

Reference [001] [011] [111] [001/[l11] [001]/[011]
[56] 1050 297 145 7.2 3.5
[57] 981 294 147 6.7 3.3
[7] 1246 245 --- --- 5.1
[18] 1385 195 --- --- 7.1
[59] --- 217 264 5.25 ---
[48] 497 120 --- --- 4.1


Table 2-4. Ductility and Strain to Fracture Data for Single Crystal NiAl

Crystal Tensile Ductility (%) Compressive Strain
Orientation (300 K) to Fracture (300 K) BDTT (K)
[001] 0 0.03 to > 0.15 575-675
[011] 0.5 to 2.5 0.10 to > 0.50 475
[111] 0.5 to 2.5 0.12 to 0.17 475
[123] 0.7 0.06 to 0.36 450









Ductility. Contrary to the behavior of many metals, single crystals of NiAl display
very limited ductility in tension at room temperature, although considerable strain to

fracture is possible in compression. Measurable tensile ductility has not been observed in

the "hard" [001] orientation. Some results of several single crystal studies [52, 59, 60, 65]

are summarized in Table 2-4. Interestingly, superplastic-like behavior has been observed

in [110] oriented crystals at 700K and 10-4s-1 strain rate in which 200 percent elongation

was achieved and "knife edge" fractures occurred [52].

These data bring to light a major concern, that the Peierls stress is very near to the

fracture stress at low temperatures where thermal activation contributes only minimally to

dislocation motion. Thus, the only way to increase the intrinsic toughness may be to

decrease bond covalency and the degree of ordering, preferably without detracting

significantly from the melting point or density.


Deformation and Fracture

The previous section was intended to give the reader an introduction to the

mechanical behavior of NiAl at room temperature. The large anisotropy of the mechanical

properties is a direct result of the operative slip systems and their relative difficulty of

activation. This section on deformation and fracture will attempt to explain some of the

controlling factors which determine the operative slip systems and their implications with

respect to strength, ductility and fracture.


Plastic Deformation and Slip Systems

Plastic deformation, in the strict sense for crystalline materials, takes place by the

motion of dislocations along particular directions and planes. For NiAl at ambient

conditions there have been several studies, using both transmission electron microscopy

(TEM) and surface slip trace analysis (SSTA), which have determined the operative slip








system to be <001>{ 110) for most cases [8, 9, 10, 57, 59, 60, 66, 67, 68, 69]. These are
shown in Table 2-5. There are two special cases where the slip system differs. [100]
oriented crystals, which have (ideally) no resolved shear stress on any <001>{110}
system, may slip on <111> 110), <111>[112}, or <110>(1TO) [10, 59, 67, 68]. [110]
oriented crystals appear to slip on {100) planes, which have a higher Schmid factor than
{ 110) in this orientation [57, 60, 66, 68]. Only one detailed study of room-temperature
deformed polycrystalline material (extruded powder) was noted, that of Bowman et al. [69].
Since extruded NiAl is known to develop a <111> fiber texture during extrusion [8, 42,
51], it might be expected that slip on the 100) plane would be quite rare, since in [111]
single crystals slip occurs on the {110) plane.

Table 2-5. Observed Slip Systems in NiAl Deformed at 300K

Reference Slip System Method Crystal Orientation
[56] <001>(110) SSTA [111]
[7] <001>{110) SSTA [111]
[67] <001>1110} TEM [111]
[59] <001>{110} TEM [111]
[48] <001>{110) TEM [110]
[56] <001>{100) SSTA [110]
[7] <001>{100} SSTA [110]
[67] <001>{ 100) TEM [110]
[59] <001>{100) TEM [110]
[67] <011>(OT1} TEM [100]
[10] <111>{???) TEM [100]
[66] <111>{110) TEM [100]
[8] <001>{???} TEM polycrystalline
[68] <001>{110) TEM polycrystalline

The operation of <001> slip is not common to all B2 compounds; in a number of
others, such as FeAl or CuZn, <111> slip is preferred [61]. The question of what factors
are important in determining the slip system in B2 compounds was initially examined by
Rachinger and Cottrell [64] who noted that highly ordered compounds of primarily ionic
bonding displayed <001> slip while those with largely metallic bonding displayed <111>









slip. According to [64], perfect [111] or [011] dislocations, if present, should dissociate

into unit dislocations according to


a[11l] ---> a[011] + a[100]

a[011] ---> a[010] + a[001]


According to Frank's rule (the elastic energy of a dislocation is proportional to the
square of the Burgers vector, Eel a gib2), there is no reduction in elastic energy in the above

reactions. Therefore, any imposed stress with a component containing a unit dislocation

will cause that component to move independently and dissociate the original dislocation.

For <111> slip to occur, dissociation according to


a[lll] ---> a/2[111] + a/2[111]


must take place. This reaction is accompanied by a reduction in elastic energy. The

resulting antiphase boundary (APB) which must form between the two partial dislocations

prevents <100> or <110> slip as long as the separation is significant, say the length of the

Burgers vector. This model allows an estimate of a critical APB energy, above which

<100> slip occurs, and below which <111> slip occurs, by the formula


Yc = aL.b


where, g. is the shear modulus, ao. is the theoretical shear strength (a 1/30), and b is the

Burgers vector. With typical values, ye is about 250 ergs/cm2. By the nearest neighbor

approximation, this equates to an ordering energy of about -0.06 eV, which may be

considered the critical ordering energy separating the two slip behaviors. Rachinger and

Cottrell found reasonable agreement with experiment for the compounds investigated;









however, Ball and Smallman [9] indicated that the above approach did not accurately

predict the slip systems for NiAl or AuZn. Using the Bragg-Williams approximation for

ordering energy, kTc4, and assuming that disordering occurs at the melting point of the

compound, they calculated an ordering energy for NiAl of -0.04 eV. Since this value is

less than the "critical" value determined by Rachinger and Cottrell, NiAl is predicted to

exhibit <111> slip, which is incorrect. Therefore, Ball and Smallman addressed the

problem of slip system prediction by using anisotropic elasticity theory to calculate the

elastic energy and relative mobility of several dislocation configurations. Configurations

with the lowest energy and highest mobility are expected to be most likely to operate. The

elastic energy per unit length of dislocation, E, was determined by


E = kb2 n(r
4;r r0)



where r and ro are the outer and inner cutoff radii, b is the Burgers vector and k is a

function of the elastic moduli and the line direction of the dislocation. The mobility*, S,

was estimated by


4S=e( -2ir
b b


where C is the dislocation width. C/b was estimated by


kd
b 2cb


* "Mobility" is a misnomer, since higher values of S are considered less mobile [62].








where c is the shear modulus in the slip direction and d is the spacing between glide planes.

That is, mobility increases with glide plane smoothness and ease of shear. Some results of

Ball and Smallman's calculations are presented in Table 2-6.


Table 2-6. Elastic Energy and Mobility of Several Dislocation Configurations

(hkll character E (ergs/cm)/104 S
111 T10 screw/edge 13.9/29.3 0.98/0.42
110 1TO screw/edge 10.3/15.3 0.32/0.06
100 011 screw/edge 9.3/8.8 0.46/0.51
100 010 screw/edge 9.3/7.7 0.67/0.71


From these results it is obvious that dislocations with <100> Burgers vectors are of

the lowest elastic energy and are therefore the most favorable to form. The <100>{011)

and <100>{010) systems are competitive in this respect, although <100>{011) appears to

be the more mobile of the two, and due to multiplication processes may be present in larger

proportions. Regardless of slip plane, the <100> dislocations of edge or mixed [70]

orientation are predicted to form more easily than the screw orientation. Slip by a/2

partial dislocations have not been considered in the above calculations because, as shown

by both Ball and Smallman [9] and later by Potter [63], the total energy of such

dislocations is still greater than that of the unit <100> dislocations due to the APB fault

energy.

It is worth pointing out that all the early workers mentioned above [9, 63, 64]

assumed the critical temperature for disordering (Tc) to be equal to the melting point (Tm)

in determining the APB and ordering energy. Although this assumption seems reasonable,

since long range order obviously cannot be maintained in a liquid, there is good reason to

consider otherwise. First, from a thermodynamic point of view, the onset of melting only

indicates that the liquid has a lower free energy than the solid, not that disordering is

imminent. Second, there is experimental evidence [25] that some degree of short range









order is maintained in the liquid well above the melting point. Finally, the APB energy for

NiAl calculated by Potter [63] based on Tc = Tm for faulting on <111>{ 110) was 223

ergs/cm2. This is at strong variance with values determined both experimentally by TEM

analysis [71] (at least 500 ergs/cm2) and by first principles calculations [28, 72] (800

ergs/cm2). Thus, NiAl is much less the "borderline" case of <100> slip than originally

thought.

Lautenschlager et al. [62] also examined the problem of slip system prediction in

CsCI type compounds. They studied the effect of the atom size ratio, RA/RB, as well as the

bonding type and crystal orientation (with respect to applied stress), on the observed slip

system. Using a hard-sphere model, they determined that RA/RB has a strong effect on the

slip plane roughness as much better packing can be achieved when RA/RB tends toward

0.732. For example, the perpendicular displacement of the glide plane during <100>(011)

slip is 0.159a for RA/RB= 0.732 and 0.293a for RA = RB; for <111>{011) slip these

values are 0.209a and 0.063a, respectively. Therefore, <111> slip is promoted when RA

and RB are similar. The effect of bonding type (ionic, covalent or metallic) was observed

to promote <111> slip when metallic bonding dominates and <001> otherwise. Crystal

orientation, relative to the applied stress, affects the preferred slip system according to

Schmid factor considerations, and thereby promotes <111> Burgers vectors in general.

The orientation effect is illustrated in Figure 2-10, after Lautenschlager [62].

For compounds such as NiAl, according to Lautenschlager et al., the bonding and

orientation factors tend to cancel one another, and relegate the slip system choice to the

RA/RB ratio. The estimation of the amount of local strain at the glide plane by a hard

sphere model for each slip system provides criteria for predicting the most likely system to

operate. By this approach, compounds with RA/RB tending to 1.000 will prefer <111>

slip and those with RA/RB tending to 0.732 will prefer <100> slip, with the critical RA/RB

ratio being about 0.95. For NiAl, RA/RB was estimated to be 0.847, which predicts

<100>{011), the experimentally observed system. Of course, accurate determination of









RA/RB for any intermetallic compound is difficult, as it depends on the known values for
the atomic radii and the degree of charge depletion or accumulation. Further, the charge
distribution has no requirement of sphericity, particularly in cases of strong, directional
bonding. In fact, close examination of atom positions during <100>{011) slip of CsCl
compounds reveals that this type of slip tends to maximize the continuous contact between
unlike atoms.


(101) {112)
<111> 0 <





<111> <101




(101) (001)
<0102 <010




Figure 2-10. Influence of Crystal Orientation on B2 Slip Systems. Each slip system is
favored in the shaded regions. After Lautenschlager et al. [62].

Implications of von Mises Criterion

The sole operation of <100>{011) slip can produce only three independent slip
systems [12]. This number is insufficient to allow general plastic flow in a polycrystalline
body, as five independent systems are required to produce the arbitrary shape change of
each grain to maintain contact with adjacent grains [13]. Therefore, even though the critical
resolved shear stress (CRSS) for <100>{011) slip is reasonably low [57], compressive









deformations as small as 1 percent have been shown to produce visible grain boundary

fissures due to poor slip transfer to adjacent grains [69]. In tension, these fissures provide

an easy path for intergranular fracture. This process is illustrated schematically in Figure

2-11. The lack of sufficient slip systems in NiAl has prompted research [14, 15, 16, 73] to

alter the primary slip system from <100>{011) to <111>{110), which has the required

number of independent slip systems.

The combined operation of two slip system families can also provide five

independent slip systems, when they cannot individually [74]. For example, combined

operation of <100>{011) (three systems) and <110>{1TO} (two systems) produces five

independent systems. This provides an alternative for satisfying the von Mises criterion

without <111> slip, which would otherwise require a significant reduction in ordering

energy.


intergranular
crack forms A


Figure 2-11. Effect of Insufficient Number of Independent Slip Systems on Plastic
Deformation in Polycrystals










Fracture of NiAl

The tensile fracture mode of polycrystalline NiAl is intergranular (IG) at the

stoichiometric composition [4, 49] and transgranular (TG) in off-stoichiometric

compositions [15]. As described above, IG fracture in NiAl is a natural result of plastic

flow on three independent slip systems with no other deformation mechanisms. This is

due to a shape incompatibility between adjacent grains and the creation of IG cracks which

coalesce and lead to catastrophic IG failure. However, for certain crack plane/grain

boundary orientations, it is not unreasonable that TG failure (cleavage) will be favored.

The TG failure in off-stoichiometric compositions is probably due to the difficulty of

dislocation motion past constitutional point defects. The preferred TG cleavage plane has

been reported to be (110) in stoichiometric NiAl [75].

Other arguments have been put forth to explain the presence of IG fracture in

stoichiometric NiAl. The effect of certain embrittling agents, such as phosphorus or sulfur,

which segregate to grain boundaries, is well-documented in ferrous metallurgy literature

and has been considered to explain IG fracture in NiAl. Westbrook [76] investigated the

effect of oxygen and/or nitrogen on the grain boundary hardness of AgMg and NiAl and

concluded that these interstitials segregated to the grain boundary regions and increased the

BDTT. However, more recent work by Zeller et al. [77] and George et al. [35] utilizing

scanning Auger microscopy (SAM) has shown that grain boundaries in binary NiAl are

free of any segregated impurities, including carbon, oxygen, boron or sulphur. This has

been interpreted [35] as an indication that the grain boundaries in NiAl are intrinsically

weak. In contrast, grain boundary structure simulations [78] seem to indicate that the

boundaries should not be intrinsically brittle, provided there are no excess aluminum atoms

at the boundary. Since off-stoichiometric NiAl alloys tend to fracture in a TG mode [15], it

appears that grain boundaries are as strong as the bulk. Therefore, one would conclude that









the grain boundaries in NiAl are not intrinsically (structurally) weak, but become the site of

Griffith defects when slip is activated.


Ternary Alloying Effects on Plastic Deformation

This section reviews the effects of ternary alloying additions to NiAl on plastic

deformation and mechanical properties. Primarily single phase B2 alloys are considered.

For convenience, microalloying (less than one atomic percent) and macroalloying additions

are discussed separately.


Microalloying Additions


Boron. Boron has been found to prevent grain boundary fracture in Ni3Al and to

induce ductility in doing so. Based on this work, boron was also added to NiAl in hopes

of achieving the same result [15, 35]. Law and Blackburn [15] added 0.25 percent boron

to stoichiometric NiAl in both cast/extruded and powder/extruded forms and found no

increase in ductility. In fact, the BDTT was elevated nearly 200K by the boron additions.

The fracture mode was reported to shift from IG to TG, although whether this was due to

an increase in grain boundary cohesiveness or the lack of plastic flow due to solid solution

strengthening is not known. In a later study, George and Liu [35] added 30, 100 and 300

wppm (parts per million by weight) to cast and extruded NiAl alloys and compared them
with binary NiAl. They found that boron did not impart any ductility to NiAl, although the

fracture mode was reported to be TG with as little as 30 wppm while the percent elongation

remained about two percent. George and Liu explained the change in fracture mode as

being due to grain boundary strengthening by the boron, and did indeed find a significant

segregation of boron to the grain boundaries. The boron additions also dramatically

increased the strength of the alloys, about 4500 MPa/atomic percent boron. Similar

ductility between the binary and NiAl+30 wppm alloys seems to indicate that the change in









fracture mode was due to a grain boundary effect and not simple yield strength elevation

(lack of plastic flow). Strengthening of the grain boundaries by boron would make IG

cracking more difficult, but would not necessarily facilitate slip transferral to adjacent

grains or activate additional slip systems. Thus, NiAl with 30 wppm boron probably has

smaller Griffith defects due to unaccommodated slip at the grain boundary. If the grain

boundaries are truly strengthened, further cracking should take place along cleavage planes

to produce the observed TG fractures. That is, the observation of grain boundary

strengthening by boron addition does not imply that the boundaries are inherently brittle.

Carbon. Carbon is a common impurity in most commercial purity elemental nickel

and, therefore, is probably present in nearly all NiAl alloys to some degree. George and

Liu [35] intentionally added 300 wppm carbon to "high purity" NiAl and noted a strength

increase similar to that of boron; no tensile elongation was measurable. The fracture mode

for the carbon-containing alloy was IG and SAM analysis of the grain boundary surfaces

showed that carbon does not segregate to the grain boundaries. (Recent re-analysis [79] of

the published fractographs for this alloy by point counting has shown it to be mostly TG.)

Beryllium. In the same study, George and Liu also measured the effect of adding

500 wppm beryllium to NiAl. This alloy exhibited about 3 percent tensile elongation at

room temperature and a slightly elevated yield strength (-15 percent greater than binary

NiAl). Fracture occurred by IG separation and no beryllium appeared to segregate to the

grain boundaries, which indicates that beryllium has little effect upon slip in NiAl aside

from a small solid solution hardening effect.

Chromium. Zirconium and Titanium. These elements are known to be effective

"getters" for interstitial elements such as carbon or oxygen. Field et al.. [18] added 0.05

and 0.2 atomic percent chromium to [001] oriented single crystals in an effort to measure

the effect of chromium on dislocation type and character during compressive deformation.

The chromium appeared to promote <111> slip relative to the binary; however, no APBs

in the <111> dislocations were visible by TEM and thus no significant reductions in the









APB energy were apparent. As shown above, <111> slip in binary NiAI is the preferred

slip system in this orientation in the absence of kinking, regardless of the presence of

chromium, which makes any conclusions concerning slip somewhat tentative. It was

suggested by Field et al. [18] that the ease of <111> dislocation activation was due to a

gettering effect of the chromium, since Cr7C3 carbide precipitates were observed during

TEM examination. To further investigate this effect, another NiAl alloy containing 0.45

percent Ti and 0.05 percent Zr (two very potent carbon scavengers) was prepared. This

alloy behaved similarly to the chromium-containing alloys, although much greater

strengthening was observed, perhaps due to the larger size of the Zr atoms. No solute

softening was apparent, which may indicate a lack of gettering since the elements

responsible for solid solution strengthening are removed from solution to form carbides.

Therefore, the enhanced <111> activation was considered to be due to "differential

proportional hardening," a general strengthening of all slip systems which decreases the

relative differences in strength.

Zirconium was added to polycrystalline (cast and extruded) NiAl in a study by

Bowman et al. [69] at the 0.05 percent (500 ppm) level. In this study, Zr was found to

elevate the BDTT by about 300K and increase the strength significantly. No tensile

ductility was observed and the fracture mode in the Zr-doped alloys was a combination of

IG and TG. Extensive TEM examination determined that the operative slip system in both

binary and Zr-containing alloys was <001>{ 110). Therefore, Zr appears to have no effect

on the operative slip system in NiAl.

Iron. Gallium and Molybdenum. A recent study by Darolia [3] indicated that 0.25

percent Fe increases the tensile ductility of [110] single crystals from two to nearly six

percent at room temperature. Similar effects of 0.1 percent Ga and Mo were also

measured, in which the ductilities were 4.5 and 2 percent, respectively. As shown in Table

2-4, typically one to two percent tensile elongation is measured in single crystals of NiAl in

this orientation, which by comparison makes the results of Darolia quite striking. Since the









beneficial effect of these elements disappears at higher alloying levels, the mechanisms)

responsible is unknown at this time. However, since the slip system remains unchanged

from <001>{ 110) [80] it is possible that the solute atoms assist in dislocation nucleation.

This effect would likely increase the homogeneity of slip and therefore distribute

deformation more evenly throughout the crystal instead of to confined slip bands.

Macroalloying Additions

To effect a change in the basic plastic deformation behavior of NiAl by adding a

ternary addition, the addition must be readily soluble. The phase equilibria of a number of

elements may be conveniently grouped according to the ternary element's position in the

periodic table, Figure 2-12.




Sc Ti V M Fe Ni Z

Y Z %T c u P

LaH Ta W O IrHT Pt

SAl-Ni-X ternary phase(s)
NiAl-X pseudobinary eutectic
1 High solubility in NiAl and/or forms B2 aluminide

Figure 2-12. Portion of Periodic Table of the Elements Illustrating General Alloying
Behavior of Ternary Additions to NiAl


The elements in Groups IIIB, IVB and VB scandiumm, titanium, vanadium,

yttrium, zirconium, niobium, lanthanum, hafnium and tantalum) form at least one ternary

intermetallic compound with nickel and aluminum, usually the Heusler phase, Ni2AlX,

and are usually of limited solubility [81, 82, 83, 84, 85, 86, 87, 88]. Elements in Group

VIIB (chromium, molybdenum, tungsten), and also rhenium, form "pseudobinary"









eutectic systems with NiAl and also have limited solubility in NiAl [16, 17, 84, 89, 90,91,

92]. Vanadium also forms a pseudobinary eutectic with NiAl [84, 93], and is unique in

that it is the only element known to fit into the first and second categories. The

pseudobinary eutectic systems allow two phase equilibrium between NiAl and the BCC

element and have potential for in situ ductile phase reinforcement. In fact, considerable

research has gone into the development of directionally solidified pseudobinary eutectics

such as NiAl-Cr and NiAl-Mo in which rods of alpha-Cr or alpha-Mo are distributed in a
P-NiAl matrix [94]. The Group VIII elements (plus manganese and copper) display large

solubility in NiAl and their B2 aluminides are often isostructural [73, 88, 95, 96, 97, 98,

99, 100, 101, 102]. This latter category offers considerable alloying potential.

Ion. Several studies have been conducted on single phase ternary NiAl+Fe alloys

[15, 73, 103, 104, 105, 106, 107]. Iron additions are perhaps the most intuitively obvious,

since FeAl can be ductile at room temperature, exhibits <111> slip [108], and is

isostructural with NiAl [100]. Law and Blackburn [15] added from 10 to 30 percent iron

to NiAl alloys containing a constant 48.5 percent aluminum and concluded the alloys were

too brittle for further testing based on cracks observed during machining. Patrick et al. [73]

produced a series of ternary alloys along the compositional tie-line from Ni-40A1 to Fe-

40A1, deformed them in compression at room temperature, and analyzed the dislocations

by TEM. Their results indicated that more than 30 percent iron was necessary to effect a

change in Burgers vector from <100> to <111>. In 1984, Inoue et al [107] reported a

tensile ductility of five percent for single phase Ni-30A1-20Fe wire produced by rapid
solidification processing (RSP). RSP induced a martensitic reaction, B2 -4 L20, and a

very fine grain size of 4 microns. In the same study, Ni-30A1-30Fe processed in the same

manner exhibited nil ductility. Two recent studies [103, 106] of Ni-30A1-20Fe
demonstrated that cast and extruded material with a 25 pm grain size has essentially zero

ductility and slips by the motion of <100> dislocations. In contrast, Kostrubanic et al.

[104, 105] has shown that the addition of a fine dispersion of Y203 particles to Ni-35A1-









20Fe RSP ribbon by mechanical alloying and hot pressing can produce ambient toughness

values as high as 34 MPa/m0-5, as determined by four-point bending of notched specimens

(ASTM E399). This is a substantial improvement over binary stoichiometric NiAl which

has a toughness of about 6-8 MPa/m0-5 [74, 98] and was considered to be due to the very

fine grain size and induced plastic flow [105].

Chromium. Manganese and Gallium. During the course of a study of the effect of

alloying on the microstructure and mechanical properties of cast polycrystalline NiAl-base

alloys, Law and Blackburn [15] noted four compositions which were damage-tolerant

enough to withstand electrostatic discharge machining (EDM) without extensive cracking:

Ni-48.5A1, Ni-48.5A1-5Cr, Ni-44A1-5Mn and Ni-48A1-0.5Ga. After deforming

compression specimens of these alloys to about 2 percent strain at room temperature, TEM

analysis was conducted. Both the chromium- and manganese-containing alloys were

determined to contain dislocations with <111> Burgers vectors. Slip in Ni-48.5A1-5Cr

was reported to occur on (112) planes within well-defined slip bands among spherical

alpha-chromium precipitates. These dislocations were reportedly of screw character and

were arranged in dipoles within the bands. In Ni-44A1-5Mn, <111>{112) slip was

reported although the dislocations were not confined to slip bands. In this case, the

dislocations were primarily edge in nature. Analysis of the binary and gallium-containing

alloys deformed at 300K revealed only <100>{01 1) slip. Some of the results of Law and

Blackburn are summarized in Table 2-7. In spite of the reported change in slip system for

the chromium- and manganese-containing alloys, no tensile ductility was apparent at 300K.

The low mobility of <111> dislocations was offered as a possible explanation for this lack

of ductility.

In a related study [16], directionally-solidified Ni-43A1-5Cr was also reported to

exhibit <111> slip. However, since the growth direction of the ingot was [101], as was as

the compression axis of the machined specimens, the resolved shear stress on the

<100>{011) systems were low compared with those on the <111>{ 112) systems. In this









manner, <100> slip was discouraged by at least a factor of two, based on Schmid factor

comparisons of the two most favored <100> and <111> containing systems. Although it

can be argued that <111> slip was made less difficult (recall that a factor of six was noted

above for binary NiAl), no actual preference for <111> slip over <100> slip was

demonstrated.


Table 2-7. TEM Results of Law and Blackburn [15] for Polycrystalline NiAl-base Ternary
Alloys Deformed at 300K

Composition (at.%) Slip System Character
Ni-48.5A1 <001>{ 110) edge
Ni-48Al-0.5Ga <001>{110) edge
Ni-46.1Al-2.4Ga <001>{ 110) edge
Ni-48.5A1-5.2Cr { 1T2 screw
Ni-43.7A1-4.9Mn {1T2} edge


Field et al [18] evaluated the effect of chromium additions on slip systems and

mechanical properties in [100] oriented single crystals at room-temperature. As noted

above, [100] oriented crystals slip by <111> dislocations in binary NiAl in the absence of

kinking. In Ni-49A1-6Cr and Ni-48A1-2Cr alloys, substantial precipitation of alpha-

chromium was observed. Chromium and NiAl have very similar lattice parameters, and

fine-scale solid state precipitation produces semicoherent interfaces which contain

interfacial nets of dislocations to accommodate the strain [109, 110]. These nets appear to

be of two types, square and hexagonal. Walter and Cline [109] determined the square nets

to consist of orthogonal <100> dislocations. Field et al. [18] showed the hexagonal nets to

consist of <100> and 1/2<111> dislocations and speculated that these arrays may serve as

sources for <111> dislocations. Although <111> dislocations were observed in the

microstructure, no conclusive evidence was apparent to support this idea. Further

investigation of the slip bands of <111> dislocations indicated they were being generated

from Cr7C3 precipitates, which have an orientation relationship with the matrix of









[100]ppt II [011]matrix

(001)ppt II (11)matrix


This implied that <111> dislocations were more easily generated at the

carbide/matrix interface than other sources, but says little about true alteration of the slip

system from that observed in binary NiAl. The occurrence of the Cr7C3 phase is curious,

since Cr23C6 is the carbide phase in equilibrium with alpha-chromium in a binary system

[111]. This may be due to the relative ease of nucleation of the Cr7C3 phase for the above

orientation relationship. In the same study, the results of alloys containing lower levels of

chromium indicated that a general solid solution strengthening makes <111> slip relatively

more favorable than in the binary. No definitive evidence of the preference of <100> slip

over <111> slip in any orientation other than [100] was shown.

Other Elements. In addition to those already mentioned, Law and Blackburn [15]

also examined the effects of additions of cobalt, beryllium, titanium, niobium, tantalum,

silicon, vanadium and zirconium. However, none of the ternary alloys containing these

additions was reported to demonstrate any ductility at room temperature, nor were any

detailed examinations of slip phenomena performed. Darolia et al. [14] studied the effect

of up to 10 percent vanadium, substituted for aluminum, on the slip and mechanical

properties in [100] oriented single crystals. They determined that vanadium causes

significant solid solution hardening and also noted the presence of Ni2AlV (Heusler, L21)

precipitates which are expected to produce further strengthening. None of the alloys

exhibited measurable room-temperature ductility and TEM analysis revealed only <100>

dislocations.









Prediction of Alloying Effects on Plastic Deformation of NiAl

Generally, two approaches to predicting the effect of alloying on plastic
deformation have been utilized, semi-empirical and theoretical. Both acknowledge the high

ordering energy of NiAl, which is directly related to the high APB energy. These high

energies promote nickel-aluminum bonds and likewise resist like-pair bonds, which is the

reason for the difficulty of <111> { 110) slip and ultimately, the reason for the operation of

only three independent slip systems. These approaches endeavor to make <111> slip

easier by substituting elements on the B2 NiAl lattice which will lower the ordering energy

of the compound.

Phenomenological Approaches. The calculated and measured ordering energies for

a number of B2 aluminides show a steady reduction in magnitude [112, 113]


NiAl > CoAl > FeAl > MnAl > CrAl > VAl


Therefore, it is logical to add elements such as vanadium, chromium or manganese to

NiAl, substituted for nickel, to try to promote <111> slip. The relationship between

ordering energy (or heat of formation) and slip system has been demonstrated by Cotton et

al. [114], in which lower ordering energies correspond to <111> slip while higher ordering

energies correspond to <100> slip, Figure 2-13. Another semi-empirical approach is to

choose atoms for substitution based on their respective size and alter the mean atomic radii

ratio (RNi/RAI). Then, as described by Lautenschlager et al. [62], when RNi/RA1 tends to

1.000, <111> slip will be preferred. For instance, aluminum is a much larger atom than

nickel. To bring RNi/RAI nearer to 1.000, one should substitute larger atoms for nickel.

Law and Blackburn [15] considered both of these approaches in devising their alloying

schemes. In the latter case, they added Be for aluminum in an attempt to bring the mean









RNi/RAI nearer to 1.000. As already noted, none of the alloys investigated by Law and

Blackburn displayed any measurable tensile ductility at room temperature, probably

because of solid solution hardening.

Theoretical-Computational Approaches. Quite recently, first-principles calculations

based on interatomic potentials, combined with the use of high-speed supercomputers,

have allowed the a priori determination of material properties. Calculations for materials

for which some properties are already known, such as NiAl, allow a check for accuracy.

Hong and Freeman [29, 72, 115] employed all-electron self-consistent total-energy linear-

muffin-tin-orbital calculations using a super cell approach to determine the 1/2<111>{ 110)

and 1/2<111>{ 112) APB energies for binary and ternary NiAl alloys. For binary

stoichiometric NiAl they calculated energies of about 800 mJm-2 for both cases, a result

which is consistent with the experimental results of Vessyiere and Noebe [71] The

substitution of chromium for aluminum, or vanadium for nickel, is predicted to decrease

the calculated APB energy to 250 mJm-2. However, the supercell models employed in

these calculations assumed 17 percent solubility; as already mentioned, the solubilities for

these elements are quite low (less than two percent at 300K). Fu and Yoo [28] also

determined the <111>{ 110] APB energy for binary NiAl using total energy calculations

and arrived at a value of 810 mJm-2. Similar to Hong and Freeman, they also suggested

that substitutions of less electropositive 3d transition elements for aluminum would

decrease the APB energy. In unpublished work, Hahn and Vedula [116] examined the

effect of substitution of a large number of ternary elements on the calculated ordering

energy, using the pseudopotential method. Their results indicated that the substitution of

rhenium, vanadium or silicon for nickel should decrease the ordering energy of NiAl.

The phenomenological and theoretical models discussed above ignore the effect of

solid solution hardening. For a substitutional element to effect a basic change in the

bonding of NiAl, it is reasonable that substantial amounts must be soluble, based on the

information shown in Figure 2-13. As a first approximation, a rule-of-mixtures may be









applied to estimate the amount of solute required to place the alloyed NiAl compound

within the region of <111> slip. This line of reasoning would require at least half the

nickel to be replaced with another transition element, a substantial amount. If such

additions cause the yield strength to be increased above the fracture strength by solid

solution hardening before the ordering energy is effectively decreased, the alloy will not be

able to undergo any significant deformation prior to fracture and their purpose will be

defeated.


S<111> slip NiA
NiAl
O <100> slip O
A metastable
PdIn
60 0 CoAl
0


----------------^ -----
FeAl

AuZn
V MnAl O
VAl A
20 CrAl A AgMg
A NiZn
CuZn
AgCd AgZn
I I -- .


In (disorder parameter)



Figure 2-13. Relationship Between Heat of Formation, Degree of Disorder and Observed
Slip System at 300K. Note that higher order and larger heats of formation correspond to
<100> slip [114].


Another important factor which has not been considered in current theoretical

models is site preference of the ternary addition. If the alloying addition is substituted





36


compositionally for a particular site (say nickel), but instead prefers the other site (say

aluminum), then the intended effect has not been realized. Further, such a case would lead

to nickel vacancies and the associated hardening. Therefore, not only must the

substitutional addition decrease the ordering energy, it must be adequately soluble, an

impotent hardener, and exhibit the proper site preference in the NiAl lattice.














CHAPTER 3
EXPERIMENTAL PROCEDURE


The materials used in this study were obtained in several forms: homogenized arc-

melted (HAM) castings, as-arc-melted (AM) castings, extruded vacuum-induction-melted

(XVIM) castings and extruded atomized powder (XAP). The different production

methods are described separately below. The experimental work was performed while in

residence at the NASA Lewis Research Center (LRC) in Cleveland, Ohio. Compositions

are given in atomic percent, e.g. Ni-45A1-5Cr indicates 45 atomic percent aluminum, 5

atomic percent chromium and 50 atomic percent Ni.


Materials


HAM and AM Castings


The nominal and analyzed compositions of the 15 castings used in this study are

plotted in Figure 3-1. The castings were produced by nonconsumable arc melting 0.070 kg

of constituent elements (99.95 percent nickel, 99.999 percent aluminum, 99.99 percent

chromium, by weight) in a water-cooled copper bowl under argon to produce buttons,

Figure 3-2, and then inverting and remelting twice more to insure homogeneity.

Subsequent weighing of the buttons indicated weight losses were generally less than 0.1

percent. The buttons were cast into a copper mold of approximate dimensions 50 mm x

13 mm x 13 mm to produce somewhat rectangular castings with a small "buttonhead" to

reduce pipe shrinkage. The castings were subsequently homogenized by heating under

slow-flowing commercial purity argon for 24 hours at 14000C, utilizing heating and

cooling rates of 5K per minute. The AM castings were analyzed in the as-cast state.









XVIM and XAP Extrusions

Five compositions were produced by this processing route, to coincide with five of
the HAM casting compositions. 1 kg charges of constituent elements were melted together
under vacuum by induction heating to produce 0.15 m by 0.05 m diameter castings. One
composition, Ni-45Al-5Cr, was vacuum atomized to produce powder by Homogeneous
Metals, Inc. The castings and resulting atomized powder were enclosed in 1018 steel cans
by tungsten-inert-gas welding which were subsequently evacuated and sealed by electron
beam welding. Each canned alloy was heated at 11270C for four hours prior to extrusion
in a Loewy vertical hydraulic 340 ton press at an extrusion ratio of 16:1.


Al Arc Melted and Drop Cast
55 0 + 1400'C/24 Hours
0 Pratt & Whitney Alloy 40
O Vacuum Induction Cast and
Extruded
LO Extruded Powder Alloy



Al-rich
%A1 .... % Cr





SNi/Al=1
Ni-rich V

45 M 10
Ni 55 50 45 Cr
%Ni


Figure 3-1. Ternary Diagram of Alloy Compositions in Current Study









Prior to further investigation, the extrusions were cut into 0.05 m lengths and the

steel can removed chemically by immersion in an aqueous solution of 50 percent nitric acid

with a small amount of sulfuric acid for 1 hour at room temperature.


Other Materials

During the course of the investigation it was necessary to obtain some alloys

which, although not an intrinsic part of the original study, were necessary to resolve

particular questions. These include the following: (1) a portion of a vacuum-induction-

melted Ni-48.5A1-4.2Cr casting produced at United Technologies Pratt and Whitney

which was originally investigated in reference [15] (courtesy Dr. S. Russell); (2) three

NiAl single crystals in [100], [111] and [123] orientations from LRC (courtesy

R.D. Noebe) for microhardness determination only; (3) HAM Ni-43A1 and Ni-43A1-5Cr

castings in response to discussions with D.B. Miracle and his work referenced in [16]; (4)

two series of five HAM castings containing 0, 0.25, 0.50, 0.75 and 1 atomic percent

chromium (Ni/Al ratio = 1), to establish the effect of constituent nickel reported to contain

very low interstitial levels (high-purity nickel provided courtesy of Dr. S. Chumley, Ames

Research Laboratory, Ames, Iowa) and (5) seven other additional HAM castings for

microhardness determination only, to establish trends indicated by the original 15

compositions described above.


Optical Microscopy

Transverse cross-sections of the castings and extrusions were removed by abrasive

wheel cutting for optical metallographic analysis and microhardness testing. The slices

were metallographically prepared used standard mounting and grinding procedures,
polished through 0.05 pmn alumina media and then etched with saturated molybdic acid















A I I I 1-1-A I I I I I
U

S


Figure 3-2. Arc-Melted Button and Casting of Binary NiAl









(0.100 kg MoO3, 50 ml HF and 150 ml H20) to reveal the microstructure. Light optical

micrographs were recorded for each alloy, with the exception of the binary stoichiometric

Ni-50A1 due to its extreme resistance to etching. Both bright field (BF) and differential

interference contrast (DIC) were used to illuminate particular microstructural features.


Mechanical Property Testing

Compression Testing


Castings. The body of each casting was sent to Ultracut, Inc., West Palm Beach,

Florida, for electrostatic discharge machining (EDM) of cylindrical compression test

specimens, Figure 3-3 and 3-4. Approximately 16 compression specimens were obtained

from each casting, with the exception of Ni-52Al-lCr from which none could reliably be

produced due to its extreme brittleness. The compression specimens were specified to be

6.0 + 0.1 mm in length by 3.0 + 0.01 mm diameter, as this is a convenient diameter for

subsequent TEM specimen preparation. Following EDMing, a longitudinal burr remained

on each compression specimen, as well as a gray oxide. The burr was removed by lightly

grinding on 400 grit SiC paper; the specimen ends were briefly abraded to remove the

oxide layer. Parallelism of the ends was checked during specimen measurement by

placing the specimen in a micrometer and examining for light visibility between the

specimen and micrometer platens.

Extrusions. Compression specimens of identical dimensions were produced from

the extrusions by centerless grinding to final dimensions. This method produces a

smoother surface than that associated with EDMing. Thus, the properties derived from the

extrusions were more reproducible than those of the castings for a given alloy.

Compression tests were conducted primarily to introduce a small amount of

deformation into the alloys for the purpose of dislocation analysis. Since preliminary









Casting


25















3mm +-
0.05 mm






Rod


- 15 mm


~ 15 mm


3 mm +-
0.05 mm



6 mm +- ends must be flat
0.1 mm i and parallel to
within 1 degree



Specimen


Figure 3-3. Schematic Drawing of Compression Test Specimens





























Figure 3-4. Electrostatic Discharge Machined Compression Specimens


research indicated that strains in excess of about one percent produce dislocation densities

too high for individual dislocation analysis, the deformation strain was generally limited to

0.5 percent. In addition, at least four tests to failure were conducted for each cast and

homogenized composition to allow determination of 0.2 percent offset yield (CYS) and
ultimate compressive strengths (UCS) and strain-to-fracture (Ep). All tests were conducted

at a temperature of 300K in an Instron load frame, at a chart speed of 2.117 x 10-4 m per

sec (0.5 inches per minute) and a crosshead speed of 8.47x10-7 m per sec (0.002 inches

per minute). The nominal strain rate was 10-4 m/m. Boron nitride was used to lubricate

the specimen ends prior to each test and preload of approximately 10 kg was applied before

initiating crosshead travel. Following each test, the strain in each specimen designated for

TEM specimens was measured directly from the reduction in length by micrometer. In

specimens intended for mechanical property determination only, failure was considered to

have occurred upon observation of the first discontinuity in the load-displacement curve,









regardless of any subsequent increase in load. The test was usually terminated after

observation of the second load drop. In data reduction, only engineering stress and strain

were reported, due to the small strains-to-failure generally observed.


Tensile Testing

Buttonhead type tensile specimens of dimensions shown in Figure 3-5 were

machined from the extrusions by centerless grinding and electropolishing to remove

surface defects. The same Instron load frame used for the above compression specimens

was used for testing and the data were reduced in similar fashion. In addition, limited

elevated temperature tests were conducted to determine the brittle-to-ductile-transition-

temperature (BDTT) of each of the five extrusion alloys.


Microhardness Testing

All microhardness determinations were conducted on a Buehler Micromet II

microhardness testing machine using a Vickers indenter. The indenter load was chosen to

be 0.5 kg, with a load duration of 15 s. Prior to each testing session the machine was

calibrated using standard hardness test blocks of hardened steel. Specimens were

metallographically prepared prior to testing and either 10 or 15 indentations were made in

each, sampling at least five different grains. Each indentation was optically examined for

cracking or subsurface defects prior to measurement; any such defective indentations were

ignored. The resulting compilation of hardness values were averaged for each alloy.


Transmission Electron Microscopy (TEM)


Specimen Preparation


Compression specimens were sliced into disks 300 pgm thick with a low-speed

Buehler saw using 250 pmn thick SiC blades with water as a lubricant and then wet ground









by hand on 600 grit SiC paper to approximately 150 gtm in thickness. Each disk was then

jet-polished to perforation in a solution of 62 ml perchloric acid, 100 ml butyl cellusolve,

137 ml H20 and 700 ml ethanol, at 0C, 32 V and 150 mA using a Struers TENUPOL-3

polishing unit. The pumping speed of the jets was as slow as possible to minimize

bending damage in the thin area.


Microstructural Characterization and Dislocation Analysis

All observations were made on a JEOL 100C transmission electron microscope

(TEM) operated at an accelerating voltage of 120 kV on a double-tilt stage allowing + 600

and + 300 tilts. Both extruded and cast material were studied by TEM. Representative

micrographs were recorded and the presence of second phases were noted and identified

using electron diffraction analysis. In addition, the association of dislocations with

particular microstructural features, such as grain boundaries and precipitates, was noted.

The Burgers vectors of dislocations was determined by the "invisibility criterion" (zero

contrast) as described by Hirsch, Howie and Whelan [116] by recording bright field

images of dislocations illuminated by various diffracting vectors. In general, 25 to 100

dislocations could be imaged in one field of view at 20,000 to 50,000 magnifications (X).

Enlarged prints were produced from the images which allowed the determination of the

slip vector for a large number of dislocations within a given grain. For most compositions,

at least three different grains were analyzed in foils from at least two different specimens to

achieve a representative sample of Burgers vectors and help eliminate errors in analysis and

orientation effects. Because of the elastic anisotropy in NiAl and extended dilation around

the edge components of dislocations, some residual contrast sometimes remained in spite

of satisfying the invisibility criterion. With practice, it was possible to identify the visibility

conditions with little ambiguity. To insure the accuracy of the analyses, typically six

different imaging conditions were recorded, as opposed to the minimum of three. Selected









specimens were analyzed in further detail to determine dislocation line direction, character

and slip plane, in addition to Burgers vector, by appropriate stereographic techniques [10].


TOTAL LENGTH IS NOT
FAIRING MUST CRITICAL. STOCK IS
MEET 3.2 DIA AT ALREADY CUT TO
POINT OF LENGTH. NO FURTHER
TANGENCY WITH CUTTING NECESSARY
NO UNDERCUT


6.50
NOMINAL
DIA REF

_L


3.2 DIA + 0.08 -
DIA MUST NOT
VARY OVER 0.013
FOR 30 LENGTH


GRIND TO MAKE
BOTH ENDS ROUND

5.08 TYP


NOTE


1. ALL DIAMETERS MUST BE CONCENTRIC AND TRUE
WITHIN 0.013

2. V UNLESS OTHERWISE NOTED

3. MATERIAL IS EXTREMELY BRITTLE. HANDLE AND
MACHINE WITH CAUTION

4. ALL DIMENSIONS IN MILLIMETERS


Figure 3-5. Tensile Test Specimen Produced from Extrusions by Centerless Grinding










ALCHEMI


A determination of the site preference of chromium on the B2 NiAl lattice was

made by axial Atom Location by Channelling Enhanced Microanalysis (ALCHEMI) [117,

118], via the TEM, of extruded Ni-49.5Al-lCr along the [110] zone axis. This alloy was

chosen because the chromium was entirely in solid solution and thus, interference from

alpha-chromium precipitates could not occur. The technique is described in detail by Otten

[117], however, it basically consists of acquiring X-ray spectra while the electron beam is

aimed exactly along the [110] crystal axis (the [100] direction would do just as well). The

resulting spectrum is compared with another spectrum acquired slightly off (about two

degrees) the [110] axis along a high index direction. Because of the tendency for a

standing electron wave to develop in the [110] direction, columns of atoms of one type,

e.g. nickel, will be excited to a greater degree than the other, in this case, aluminum, and

this will be reflected in the x-ray spectrum relative to the off-axis analysis. If chromium

displays a site preference, its peak intensity will be augmented or diminished according to

the degree of site preference.


X-ray Diffraction Analysis (XRD)

The lattice parameters of the alloys were determined by XRD on a Scintag PAD V

diffractometer with powders produced by crushing samples of the castings. Copper Ka

radiation was used. Because of the similarity in the lattice parameters of the B2 and A2

phases in the two-phase alloys, deconvolution of the peaks was not always possible and

may have influenced the results.














CHAPTER 4
RESULTS

The results of the experiments described in the previous section are presented

below in similar order. Justification for the particular experiments chosen and critical

evaluation of the data are reserved for the discussion.


Alloy Compositions

The nominal and analyzed compositions of the materials in this study are

summarized in Tables 4-1, 4-2 and 4-3. Table 4-1 pertains only to HAM castings with the

exception of Ni-48.5A1-5.2Cr, which was also studied in the AM state. The analyzed

compositions of the castings appeared to indicate higher aluminum contents and slightly

lower nickel and chromium contents than intended. However, based on the very small

casting weight losses measured, and the fact that the aluminum content appears to have

increased (unlikely), it is probable that the chemical analysis is in error and the nominal

compositions are reasonable measures of the true compositions. The extrusion

compositions appear very near to their nominal compositions, Table 4-2.


Optical Microscopy

Representative light optical micrographs of the castings and extrusions are

presented below in Figures 4-1 through 4-10 and 4-11 through 4-15, respectively.


HAM Microstructures


The structures of the cast and homogenized binary alloys, Ni-48A1, Ni-50A1 and

Ni-52A1, were single phase and coarse grained, as shown in Figures 4-1 and 4-2. While









Ni-50A1 appeared to cast well and contained few defects, the other two binary alloys

contained a large number of cracks which were usually intergranular. Alloys containing

one percent chromium were generally single phase (Figure 4-3), with occasional fine

precipitation. Based on available ternary phase diagrams, any second phase precipitation in

this system is expected to be alpha-chromium. Both XRD and TEM analysis confirmed

this to be true, as discussed below. Alloy Ni-48A1-2Cr was very similar to the one percent

chromium alloys, as shown in Figure 4-4. These alloys also displayed a significant degree

of cracking and shrinkage pores. In contrast to Ni-48A1-2Cr, Figures 4-5 and 4-6 show

the microstructures of Ni-49A1-2Cr and Ni-50A1-2Cr to contain fine, intragranular

precipitation. The precipitates in these alloys were often aligned in smoothly curving arcs.

The precipitation in Ni-50A1-2Cr appeared to be of a slightly larger mean size than that in

the Ni-49A1-2Cr. The precipitation in the alloys containing five percent chromium was

marginally coarser than in the alloys containing two percent chromium, and was less

homogeneously distributed, Figures 4-7 through 4-9. The alignment of precipitates was

less pronounced in the Ni-45A1-5Cr and Ni-47.5A1-5Cr alloys, but was well-developed in

the Ni-50A1-5Cr alloy. In these latter alloys, the grain boundaries were often delineated by

allotriomorphs and a surrounding precipitate free zone, Figure 4-9. The microstructure of

the Ni-48.5A1-5.2Cr alloy was very similar to Ni-50A1-5Cr and is shown in Figure 4-10.


XVIM and XAP Microstructures


The optical microstructures of the extrusions are shown in Figures 4-11 through 4-

15. The binary Ni-50AI and the Ni-49.5Al-lCr alloys both contain fully recrystallized

structures of fairly equiaxed grains which have undergone dynamic recrystallization and

grain growth. The Ni-49A1-2Cr and Ni-48A1-2Cr alloys are similar, but have retained

partially recrystallized regions containing fine subgrains and a generally finer structure.

The highest chromium alloy, Ni-45A1-5Cr, displayed the finest grain structure and was









fully recrystallized. In all the chromium-containing alloys, stringers of second phase could

be discerned extending in the extrusion direction.


Table 4-1. Nominal and Analyzed Compositions (Atomic Percent or Wppm) of Cast and
Homogenized Alloys in Current Study


Nominal


Analyzed


% % % % % % % wpm ppmpp %Wt.
Alloy # Ni Al Cr Ni Al Cr Fe Cu C N 0 Loss
1 52 48 0 51.3 48.7 na nd nd 60 nd 74 -0.11
2 50 50 0 48.8 51.2 na nd nd 58 nd 76 -0.07
3 48 52 0 47.0 53.0 na nd nd 48 nd 60 -0.09
4 52 47 1 51.0 48.2 0.83 nd nd 50 nd 87 -0.19
5 50 49 1 48.7 50.4 0.90 nd nd 58 nd 46 -0.18
6 49.5 49.5 1 48.6 50.5 0.82 nd nd 56 nd 59 -0.32
7 49 50 1 48.2 51.0 0.82 nd nd 57 nd 25 -0.22
8 47 52 1 45.4 53.7 0.89 nd nd 56 nd 27 -0.17
9 50 48 2 48.8 49.5 1.7 nd nd 57 nd 36 -0.20
10 49 49 2 48.1 50.3 1.6 nd nd 50 35 86 -0.14
11 48 50 2 47.0 51.3 1.7 nd nd 64 31 56 -0.41
12 50 45 5 49.7 46.1 4.2 nd nd 59 10 97 -0.33
13 47.5 47.5 5 46.5 49.4 4.1 nd nd 55 nd 43 -0.20
14 45 50 5 44.1 51.7 4.7 nd nd 116 138 148 -0.24
15 46.2 48.5 5.2 44.2 50.2 4.7 0.84 nd 240 31 45 na

nd = not detected (less than 0.2 wppm)
na = not analyzed


Table 4-2. Nominal and Analyzed Compositions (Atomic Percent or Wppm) of Extruded
Alloys in Current Study


q


^-^-^-^^-^- ^-_>__-_ __--_ ^-_- S a^_-_---^
Extrusion Billet % % % % % % % wppm wpp wppm
# T Ni Al Cr Ni Al Cr Fe C N O
L-2692 VIM Cast 50 50 0 50.2 49.8 na na 82 nd 42
L-2661 VIM Cast 49.5 49.5 1 49.7 49.3 0.94 na 110 na 35
L-2663 VIM Cast 49 49 2 49.2 49.9 1.9 na 79 na 38
L-2659 VIM Cast 50 48 2 50.1 48.0 1.9 na 70 na 42
L-2614 Powder 50 45 5 50.9 44.6 4.4 0.08 21 nd 120


nd = not detected (less than 0.2 wppm)
na = not analyzed


___ __


Nominal


Analyzed








Table 4-3. Compositions of Additional Alloys (Atomic Percent or Wppm)
I -


-...--- ---
Alloy % % % % % % % ppm pp ppm
Identification Ni Al Cr Ni Al Cr Fe C N O
P&W Alloy 40 47.3 48.5 4.2 47.2 48.7 4.0 0.09 94 8 24
[100] NiAl SC 50 50 0 na na na na na na na
[111] NiAl SC 50 50 0 na na na na na na na
[123] NiAl SC 50 50 0 na na na na na na na
Ni-43A1 57 43 0 57.0 42.9 0 0.09 58 58 116
Ni-43A1-5Cr 52 43 5 52.4 43.6 3.94 nd 56 125 116
CP NiAl 50 50 0 na na na na 60 42 111
CP NiAl+0.25Cr 49.88 49.88 0.25 na na na na na na na
CP NiAl+0.50Cr 49.75 49.75 0.50 na na na na na na na
CP NiAl+0.75Cr 49.63 49.63 0.75 na na na na na na na
CP NiAl+1.0Cr 49.50 49.50 1.0 na na na na na na na
HP NiAl 50 50 0 na na na na 56 nd 116
HP NiAl+0.25Cr 49.88 49.88 0.25 na na na na na na na
HP NiAl+0.50Cr 49.75 49.75 0.50 na na na na na na na
HP NiAl+0.75Cr 49.63 49.63 0.75 na na na na na na na
HP NiAl+1.OCr 49.50 49.50 1.0 na na na na na na na
Ni-49.25Al-1.5Cr 49.25 49.25 1.5 na na na na na na na
Ni-48.25Al-3.5Cr 48.25 48.25 3.5 na na na na na na na
Ni-51.5Al-2.5Cr 46 51.5 2.5 na na na na na na na
Ni-46Al-2.5Cr 51.5 46 2.5 na na na na na na na
Ni-46.5Al-3.5Cr 50 46.5 3.5 na na na na na na na
Ni-50A1-3.5Cr 46.5 50 3.5 na na na na na na na
Ni-46A1 54 46 0 na na na na na na na

nd = not detected (less than 0.2 wppm)
na = not analyzed



Mechanical Properties



Compression Tests


HAM Castings. The room-temperature compression test results are given in Table

4-4 and Figure 4-16. Typical stress-strain curves are shown in Appendix A. Because of

the error involved in measuring the mechanical properties of cast material due to the large

grain size (relative to specimen size) and casting inhomogeneities, comparisons should be


Nominal


Analyzed









made with caution. However, several observations may be made. First, compositions for

which the nickel-to-aluminum ratio is unity correspond to a minimum in strength,

regardless of chromium content. Second, substantial strengthening (about three times)

occurs between zero and one percent chromium for all three series of alloys. An

interesting result in Figure 16b is that chromium, when substituted at low levels for nickel,

hardens NiAl in a manner similar to nickel in the binary alloy. Finally, alloys with

chromium levels beyond one percent demonstrate almost no additional strengthening when

the nickel-to-aluminum ratio is one, and moderate strengthening otherwise.


Table 4-4. Results of Compression Tests of HAM Castings at 300K (error = std. dev.)

Alloy 0.2% CYS (MPa) UCS (MPa) Plastic Strain (%) # tests
Ni-48A1 462 + 25 527 +34 0.4 + 0.3 8
Ni-S50A 148 + 29 468 171 12 + 10 8
Ni-52A1 834+ 186 867 238 0.35 0.2 7
Ni-47A1-1Cr 543 72 606 + 64 0.73 + 0.4 8
Ni-49Al-lCr 450 + 56 534 + 68 0.5 + 0.2 9
Ni-49.5Al-1Cr 403 + 100 457 +70 0.9 + 0.8 9
Ni-50AI-1Cr 461 +97 514 + 92 0.5 + 0.4 8
Ni-52A1-1Cr no data 246 no data 2
Ni-48A1-2Cr 461 49 500 49 0.5 + 0.2 8
Ni-49A1-2Cr 440 +42 560 + 49 0.7 + 0.3 9
Ni-50Al-2Cr 571 + 81 599 + 69 1.0 + 0.8 8
Ni-45A1-5Cr 753 + 69 870 + 63 0.6 + 0.1 8
Ni-47.5A1-5Cr 453 +68 608 +9 1.3 + 1.5 7
Ni-50Al-5Cr 816+ 151 885 + 94 0.2 + 0.3 7


Extrusions. The compressive yield strengths for the XVIM and XAP extrusions

are given below in Table 4-5. Values for strain-to-failure and fracture strength were not

recorded for the extrusions. Plotting the yield strengths of the castings versus the

extrusions for identical alloy compositions, Figure 4-17, shows that the relative differences

in strength between alloys is consistent for both processing routes, although the extruded

material is generally about 40 MPa stronger.






























Figure 4-1. Optical Microstructure of Ni-48A1 HAM Casting


Figure 4-2. Optical Microstructure of Ni-52A1 HAM Casting







































y '*25um"


Figure 4-3. Optical Microstructure ofNi-49Al-Cr Casting

Figure 4-3. Optical Microstructure of Ni-49AI-1Cr HAM Castig0


Figure 4-4. Optical Microstructure of Ni-48A1-2Cr HAM Casting


. ..... 7 -


,,



.
..~.






























Figure 4-5. Optical Microstructure of Ni-49A1-2Cr HAM Casting


Figure 4-6. Optical Microstructure of Ni-50A1-2Cr HAM Casting














S ," ..". ''"** ..* *.-

'*. ..* *
'. .. .*. .

"* ".-' .. *", ,' -
.'. ., .. -
a .' *. .
** *. '' *

*'i .. .* s.-
S. .

S* .*.. .. .. 1 0
o ." ". .. -
,

SV .i *. r*
.. -.
S* *, J *
SI M"

Ii' *- : .
'1 *
*;. "S. : i
.. .. ". -.
.... .. S. .


",*. ... ".. *...






*'
,S *S
















*igue *4 "tc Mi 'o -"u' oN 47A
.*~ -. ... *.a



Figure 4-7. Optical Microstructure of Ni-45A1-5Cr HAM Casting







5-,. s. .-" "3"


*..*** M.. -
,, W ,' ''" ':.t. i

.:1* .- '. .



%S :: .



.. 5> ; ..0 .1 .s. ., +,.- su






Figure 4-8. Optical Microstructure of Ni-47.5A1-5Cr HAM Casting


























:Lg *qr -: c* .
S..


..I O ~. ~,,
4AP



r r


Figure 4-9. Optical Microstructure of Ni-50A1-5Cr HAM Casting


B-4"

a
*9*
9. 8 9 *


,. 8
* .1.

'0
S


C-
-e
0I
a


V

4*..


4b .0


, .
9 4. '. .




r*. Ia' ar
~~a *r


0




* *C



-~' .
:


-,


~ .9
*
8'I P
8o


E r *


N 0 *-w0

C *v 11 IdI l
p ~ r .. -.-.--


Figure 4-10. Optical Microstructure of Ni-48.5A1-5.2Cr HAM Casting


a* 9


0


*~
9
0,


$- ..~
ar~
U

4
*'
at




















































Figure 4-11. Optical Microstructure of Ni-50A1 XVIM Extrusion (a) Longitudinal; (b)
Transverse























I


(b)

Figure 4-12. Optical Microstructure of Ni-49.5A1-1Cr XVIM Extrusion (a) Longitudinal;
(b) Transverse


AOL'




















































Figure 4-13. Optical Microstructure of Ni-48A1-2Cr XVIM Extrusion (a) Longitudinal; (b)
Transverse



















































(b)

Figure 4-14. Optical Microstructure of Ni-49A1-2Cr XVIM Extrusion (a) Longitudinal; (b)
Transverse

















































(b)

Figure 4-15. Optical Microstructure of Ni-45A1-5Cr XVIM Extrusion (a) Longitudinal; (b)
Transverse





63


1000
13 Cr for Al
A Cr for Ni
800 0 Ni/A1=1


S600


S 400 *


200

I S I I I*
0 1 2 3 4 5 6

Atom Percent Chromium
(a)


1200

1000

800- 15% Cr


S 600" 2

q 400
o 1% Cr/

200 0% Cr

0 I I I I
44 46 48 50 52 54

Atomic Percent Al
(b)

Figure 4-16. 0.2 Percent Compressive Yield Strength at 300K of HAM Castings as a
Function of (a) Chromium Content; (b) Aluminum Content










Table 4-5. Compressive Yield Strengths of Extruded NiAl+Cr Alloys at 300K

Alloy 0.2% CYS (MPa) # Tests
Ni-50AI 189 +6 5
Ni-49.5A1-1Cr 441 + 16 4
Ni-48A1-2Cr 530 +7 4
Ni-49A1-2Cr 559 + 13 4
Ni-45A1-5Cr 818 +9 7


1000


800


600


400


200


200 400 600

Cast and Homogenized 0.2% CYS (MPa)


800


Figure 4-17. Relationship Between Compressive Yield Strengths of HAM Cast and
Extruded NiAl+Cr Alloys

Tensile Tests

The yield strengths and elongations as a function of test temperature are plotted in

Figure 4-18. As in the compressive yield strengths of the HAM cast alloys, the addition of

chromium augments the room-temperature tensile strength of the extrusions about three-

fold. Less than 0.2 percent elongation was obtained for the other chromium-containing

alloys below about 700K. In general, increasing the chromium level both raises the yield









strength and increases the dependence of strength on temperature. The elongations as a

function of temperature indicate that chromium elevates the BDTT about 150K relative to

binary stoichiometric NiAl. Comparison of the tensile ductilities for Ni-49.5Al-lCr and

Ni-49A1-2Cr indicates the nickel-to-aluminum ratio has a stronger influence on the BDTT

than does the chromium content. This influence is suggested because the nickel-to-

aluminum ratio for these two alloys is equal to one while their chromium contents differ by

a factor of two; yet their BDTTs are both about 750K. Ni-48A1-2Cr and (nickel-to-

aluminum ratio equal to 1.04) and Ni-45Al-5Cr (nickel-to-aluminum ratio equal to 1.11)

display respectively higher BDTTs than Ni-49.5A1-1Cr and Ni-49A1-2Cr.


Microhardness Tests

The measurement of microhardness was considered an alternative route to

estimating yield strength variations for a number of extra compositions beyond the original

15 alloys. A reasonably good relationship between microhardness and yield strength was

initially established, as shown in Figure 4-19.

The Vickers microhardness as a function of chromium content is plotted in Figure

4-20 in three curves representing substitutions for nickel only, aluminum only, and evenly

for both. The hardening response at chromium levels below one percent is very flat and

similar for all three substitutional schemes, roughly 265 kg/mm2. At about one percent

chromium the hardness increases sharply for all three curves, the most marked change

occurring when chromium is substituted for nickel and the least marked when chromium

is substituted for both elements. Beyond one percent chromium the curves separate, with

the smallest degree of hardening occurring when chromium is substituted equally for both

nickel and aluminum, and the greatest degree of hardening occurring when substituted for

nickel. Comparison of the hardness and compressive yield strength values of the HAM









castings, Figures 4-20 and 4-16, shows that yield strength is more sensitive to property

changes than hardness.

The flat hardening response at chromium levels below one percent prompted a

study of the effect of interstitial purity levels on the hardness of low-chromium nickel

aluminide alloys. The hardness of two series of alloys containing from zero to one percent

chromium, all other variables being the same, are plotted in Figure 4-21 as a function of

chromium content. This data indicates the source of constituent nickel used to cast the

nickel aluminide alloys has a strong influence on the hardness, and probably other

mechanical properties as well. A student's t test was performed to check for the statistical

significance of the difference between the means of the hardness of the two binary alloys.

They were found to be significantly different to at least 99 percent confidence.

Interestingly, this apparent dependence on purity was not reflected in the interstitial

analyses shown in Table 4-3, except in regard to nitrogen content, an unlikely contaminant.

600

500
i49A12Cr
400
CO t Ni45A15Cr
300


200 Ni Ni49.5AllCr

100 Ni48A12Cr


200 400 600 800 1000 1200

Test Temperature (K)

Figure 4-18. Tensile Properties of Extruded Alloys as a Function of Test Temperature
(a) 0.2% Tensile Yield Strength








80

70

60
e Ni50AI
S50 Ni49.5AllCr
S 40 Ni48AI2Cr
o 304

H 20
E 20 Ni45Al5Cr

10 Ni49A12Cr


200 400 600 800 1000 1200

Test Temperature (K)

Figure 4-18--continued. (b) Tensile Elongation




1000


800 -


600

U 0
S400


200
CYS = 700.8 + 3.57Hv
RA2 = 0.889

200 300 400 500

Vickers Microhardness (kg/mm2)


Figure 4-19. Relationship Between Microhardness and CYS for HAM Cast Alloys

























0.01 0.02 0.03 0.04

Atom Fraction Chromium


0.05 0.06


Figure 4-20. Vickers Microhardness of Cast and Homogenized NiAl+Cr Alloys as a
Function of Chromium Content


Further analysis of the as-received constituent nickel did not reveal any significant

differences in interstitial or metallic impurity content. These results are in Table 4-6.


Table 4-6. Chemical Analysis of As-Received Nickel Shot (Wppm)


NASA- LRC


LECO


Nickel Ni Al Fe Cu Si C N O C N O S
Commercial bal nd nd <0.2 <0.1 97 <10 56 69 <2 57 <2
Purified bal nd nd <0.2 <0.1 84 <10 47 76 <2 42 <1

nd indicates less than 0.05 wppm


The hardness dependence upon aluminum content at constant chromium levels is

plotted in Figure 4-22 for the cast and homogenized alloys. The expected hardness

minimum at the stoichiometric composition is obvious in the binary alloy, but less


500

450

400

350

300

250


200'
0.00









pronounced in the ternary alloys which demonstrate less of an increase for nickel-rich

compositions. This is especially true of the one percent chromium alloys.

A better understanding of the overall dependence of hardness upon composition

may be achieved by a three-dimensional (3D) plot of the hardness surface as a function of

aluminum and chromium content. Such a plot is shown in Figure 4-23. The 3D plot

makes it clear that hardness minima reside along a "valley" in which the nickel-to-

aluminum ratio is equal to one. It is also shown that aluminum-rich compositions are

substantially harder than nickel-rich compositions, regardless of chromium content. A

small minimum in hardness appears to occur at the Ni-49A1-2Cr composition, however,

this minimum probably falls within the data scatter.


300


290


280


270


260


250


240 '
0.0


0.2 0.4 0.6 0.8 1.0


Atom Percent Chromium

Figure 4-21. Comparison of Hardening Responses of Low-Chromium NiAl+Cr Alloys
Produced from Commercial Purity and High Purity Nickel Sources.









500
S0O%Cr
Ct o 1%Cr
A 2%Cr
3.5%Cr
N U 5%Cr
400




S 300
I)



200 '
44 46 48 50 52 54

Atom Percent Aluminum

Figure 4-22. Vickers Microhardness of Cast and Homogenized NiAl+Cr Alloys as a
Function of Aluminum Content


Transmission Electron Microscopy



HAM and AM Castings


Typical microstructures are shown in Figures 4-24 through 4-36, while Figure 4-37

shows the microstructure of Pratt and Whitney alloy 40. All the cast alloys contained very

large grains (hundreds or thousands of microns) such that few grain boundaries were

encountered during TEM analysis. Prior to room temperature deformation, very few

dislocations were found. The deformed binary alloys were single phase, with only Ni-

50A1 containing significant numbers of dislocations which were determined to be of

<100> Burgers vectors. These dislocations were generally homogeneously distributed

after 0.3 percent deformation, and appeared to be in the form of elongated loops or short

segments. The short segments often contained one sharp bend which separated otherwise

straight lengths. An example dislocation structure of Ni-50A1 is shown in Figure 4-24.

















i....'" .~' ~~ "'i-. ....... ..



..... 358
et

S81" .


















Figure 4-23. Three-Dimensional Plot of Mficrohardness of Cast and Homogenized
NiAl+Cr Alloys as a Function of Composition
i /
.oe e,. tss,.
i/ee~ l/h II l'r'2-.,
-" o: c
`Ys-. .h
9 "" ...













Figure 4-23. Three-Dimensional Plot of Microhardness of Cast and Homogenized
NiAI+Cr Alloys as a Function of Composition









Alloys Ni-47Al-lCr and Ni-49Al-lCr were single phase. The extreme brittleness

of Ni-47Al-lCr prevented sufficient deformation to produce a workable dislocation

density. Ni-49Al-1Cr, on the other hand, contained a moderate number of dislocations and

tangles, similar to Ni-50Al. Ni-50Al-lCr contained spherical precipitates of alpha-

chromium of approximately 10 nm diameter. Occasionally, rods of alpha-chromium were

encountered of similar diameter. Both forms of precipitation were usually associated with

what appeared to be prismatically punched dislocation loops and associated loose tangles at

the precipitate/matrix interface. In the matrix, bent dislocation segments were observed

along with very small dislocation loops. The vast majority of dislocations analyzed

contained <100> Burgers vectors. An example of this microstructure and the

accompanying dislocation analysis are shown in Figure 4-25.

Ni-48A1-2Cr, Ni-49A1-2Cr and Ni-50A1-2Cr all contained spherical precipitates of

alpha-chromium, as shown in Figures 4-26 through 4-28. The precipitation in Ni-48A1-

2Cr appeared to be of a narrow size distribution centered around 100 nm, while the

precipitation in the latter two alloys tended to be bimodal, either 25 or 250 nm in diameter.

In some regions, a rod-like form of alpha-chromium was encountered. Larger precipitates

were often associated with dislocation tangles, Figures 4-27 and 4-28, while the smaller

precipitates appeared to pin matrix dislocations, Figure 4-26. The "double-arced" strain

contrast displayed by the precipitates in Figures 4-27a and 4-28a indicates the precipitates

are coherent. As for the above alloys, sharply bent dislocation segments were often

observed within the matrix, and in the Ni-48A1-2Cr and Ni-49A1-2Cr alloys, dislocation

loops were noted. All the dislocations analyzed in these alloys contained <100> Burgers

vectors. Such an analysis is shown in Figure 4-26.

The alloys containing five percent chromium, Ni-45A1-5Cr, Ni-47.5A1-5Cr and Ni-

50A1-5Cr, exhibited substantial precipitation of alpha-chromium particles, Figures 4-29

through 4-33. Precipitation in the first two alloys was generally bimodal in size, with the

finer precipitation of about 10 to 25 nm diameter and coarser precipitation of about 250 nm









diameter. While the finer precipitates were spherical, the coarser precipitates were often

irregular in morphology and were probably partially dissolved interdendritic chromium

resulting from incomplete homogenization (Figure 4-31). The alpha-chromium

precipitates in Ni-50Al-5Cr tended to be one of three different sizes: (1) fine, aligned

spheres (about 15 nm in diameter), (2) small spheres or rods (about 200 nm in diameter)

containing interfacial dislocation networks, and (3) large, aligned second phases of

approximately square cross section with rounded edges which were also associated with

interfacial misfit dislocation networks (about 1200 nm in width), Figure 4-32. The

orientation relationship between the precipitates and the matrix was determined to be cube-

on-cube, i.e., <100>a(001) a // <100>p{001) p, Figure 4-33. Pinning of matrix

dislocations by the finest precipitates was observed for all three alloys. In every case, the

Burgers vector of the dislocations was determined to be of the <100> type, an example of

which is shown for Ni-47.5A1-5Cr in Figure 4-30.

TEM analysis of the Ni-48.5A1-5.2Cr alloys was performed for both the AM and

HAM castings, and also for the material obtained from Pratt and Whitney. Although this

alloy was reported to contain 5.2 percent chromium [16], documentation obtained with the

alloy indicated it contained 4.2 percent chromium, as confirmed by chemical analysis in

Table 4-3. The AM Ni-48.5A1-5.2Cr alloy microstructure is shown in Figures 4-34 and 4-

35, and indicates that substantial segregation occurred during solidification. Figure 4-34

shows interdendritic alpha-chromium in the form of small, dense particles; identification

by convergent beam electron diffraction analysis showed these to be Cr23C6 carbides.

Away from the interdendritic regions, alpha-chromium precipitation was observed to

consist of a relatively homogeneous dispersion of fine, spherical particles, Figure 4-35.

Plastic deformation in these regions occurs in the form of concentrated slip bands of

<100> dislocations; no dislocations with <111> Burgers vectors were found.

Upon homogenization, the interdendritic alpha-chromium was eliminated and a

precipitate dispersion of a wide size distribution is produced (1 to 500 nm diameter), as









shown in Figure 4-36. These precipitates are spherical and are occasionally aligned into

short rows. Dislocation networks decorate the precipitate/matrix interface for the larger

precipitates, which are sometimes also the site of loose tangles. All precipitates appear to

participate in dislocation pinning. Unlike the AM castings, no slip bands were observed

after deformation. As in other alloys previously mentioned, most of the matrix

dislocations have the peculiar sharp bend separating otherwise straight lengths.

Figure 4-37 shows the microstructure of the Pratt and Whitney alloy (Ni-48.5A1-

4.2Cr). The precipitation in this alloy was very dense, of a wide size distribution, and was

often aligned into closely-packed rows. The microstructural homogeneity was intermediate

to the AM and HAM Ni-48.5A1-5.2Cr structures, Figures 4-34 to 4-36. No slip bands

were observed, although dislocation tangles tended to concentrate along the rows of larger

precipitates. Otherwise, the dislocation distribution was homogeneous. Some dislocation

loops were also observed. The majority of the dislocations produced by 300K deformation

were of the <100> type, contrary to published results on this material [15, 16].


Microstructures of XVIM and XAP Extrusions

In general, the extruded microstructures were fine-grained, typically 10 to 20 pm in

diameter, and displayed a <111> fiber texture which was evident during analysis of the

TEM foils. In other respects the structures differed as described below.

Ni-50AI. The as-extruded microstructures were generally featureless with the

exception of a moderately low density of <100> dislocations, Figure 4-38a. Most of these

dislocations consisted of short segments containing two straight lengths connected by a

sharp bend, as observed in the cast and homogenized material. Occasionally, small

elongated loops or partial loops were noted. Room-temperature plastic deformation (0.55

percent) increased the dislocation density, causing loose tangles, Figure 4-39; the same









sharply bent configurations were observed as above. The vast majority of the dislocations

analyzed contained <100> Burgers vectors, usually on (011) planes.

Ni-49.5Al-1Cr. This alloy was similar to the binary in the as-extruded state. No

second phase was observed and the dislocation density was relatively low. Upon room-

temperature deformation (0.6 percent), marked differences were noted in comparison to the

binary. As shown in Figures 4-40 and 4-41, deformation occurred primarily by the

propagation of concentrated slip bands with <100> Burgers vectors emanating from the

grain boundaries. These slip bands were usually planar and quite straight, however, the

intersection of two bands sometimes caused deviation, as indicated by the curved slip band

in Figure 4-41. Figure 4-40 provides evidence that the slip bands are nucleated at grain

boundaries (not merely terminating there) by the observation of half-loops extending from

the grain boundary at the initiation site of the slip bands. Dense slip bands, such as shown

in Figure 4-40, contained a considerable number of small dislocation loops with <100>

Burgers vectors. It was not clear whether the bands propagated by dislocation

multiplication or by the glide of mobile dislocations.

Ni-48A1-2Cr and Ni-49A1-2Cr. These two alloys were essentially identical in their

microstructures and dislocation substructure. Dense, homogeneous precipitation of 25 nm

diameter alpha-chromium particles were observed throughout the microstructures. Larger

Cr23C6 carbides were also commonly observed (about one or two per grain) and appeared

to serve as nucleation sites for slip bands of <100> dislocations during extrusion, as shown

in Figures 4-42 and 4-43. When deformed at room temperature, numerous parallel slip

bands were observed to nucleate at grain boundaries and extend into the grain centers,

Figures 4-44 and 4-45. Small dislocation loops, 10 to 20 nm in diameter, were usually

associated with all but the most diffuse slip bands, and appeared to be Orowan type loops

surrounding the fine precipitates. A typical Burgers vector analysis of the matrix

dislocations is shown in Figure 4-46 for Ni-48A1-2Cr.









Ni-45Al-5Cr. The as-extruded microstructure of this alloy is shown in Figure 4-47.

The alpha-chromium precipitation appears to be of a higher volume fraction, although the

typical precipitate diameter is similar to the other alloys. The dislocation density is low,

with few dislocation interactions visible. The sharp bending of short segments, obvious in

other alloys, is less prominent in this alloy. Room temperature deformation (about 0.6

percent) significantly increases the dislocation density, Figure 4-48. However, the

deformation appears more homogeneous compared to the other alloys, with few slip

bands. Dislocation pinning by the precipitates is observed, as are dislocation loops, which

occasionally surrounded the precipitates. Nearly all dislocations analyzed contained <100>

Burgers vectors.


Chromium Solubility and Site Preference in NiAl


Based on the above results, the solubility of chromium in NiAl for the given

processing conditions may be determined. A portion of the ternary nickel-aluminum-

chromium diagram is presented in Figure 4-49 with the trace of the solvus surface

indicated based on the microstructural results of the ternary alloys. This diagram indicates

that the solubility of chromium is higher for nickel-rich alloys. ALCHEMI results, taken

with the extruded Ni-49.5Al-lCr alloy along the [110] zone axis, support the

microstructural data by indicating a strong preference of chromium for the aluminum site

in NiAl. These results are plotted in Figure 4-50, which shows the intensity of the primary

K-alpha X-ray peaks as a function of deviation angle of the zone axis from the primary

beam. Since the chromium peak intensity tends to mimic the aluminum peak intensity, i.e.

it decreases with increasing deviation angle, a strong site preference for aluminum is

shown. It is not known whether nickel vacancies are created by this site preference for

alloys in which the combined chromium and aluminum contents exceed the nickel content.





















































Figure 4-24. Bright Field TEM Micrograph of <100> Dislocations in HAM Cast Ni-50A1
Deformed 0.3 Percent at 300K





















































Figure 4-25. TEM Dislocation Analysis Indicating <100> Dislocations in HAM Cast
Ni-50A1-1Cr Deformed 0.3 Percent at 300K





















































Figure 4-26. TEM Dislocation Analysis Indicating <100> Dislocations in HAM Cast
Ni-48Al-2Cr Deformed 0.6 Percent at 300K
















































(b)

Figure 4-27. TEM Microstructure of HAM Cast Ni-49A1-2Cr Deformed 0.6 Percent at
300K (a) Large Precipitates; (b) <100> Dislocations




















































(b)

Figure 4-28. Bright Field TEM Micrographs of HAM Cast Ni-50A1-2Cr Deformed 0.2
Percent at 300K


7)7-r-














(a)













(b)













(c)








Figure 4-29. TEM Microstructure of HAM Cast Ni-45A1-5 Chromium Deformed 0.3
Percent at 300K (a) Slip Bands; (b) Fine Precipitates and Dislocation Loops; (c) Coarser
Precipitates






















































Figure 4-30. TEM Dislocation Analysis Indicating <100> Dislocations in HAM Cast
Ni-47.5A1-5Cr Deformed 0.3 Percent at 300K





















































Figure 4-31. Bright Field TEM Micrographs of Large, Irregular Alpha-Chromium Phase
and Fine Precipitation in HAM Cast Ni-47.5Al-5Cr Deformed 0.3 Percent at 300K

















































(b)

Figure 4-32. Bright Field TEM Micrographs of HAM Cast Ni-50A1-5Cr Deformed 0.35
Percent at 300K (a) Pinning of <100> Dislocations by Precipitates; (b) Interfacial
Dislocation Nets at Larger Alpha-Chromium Particles






















































Figure 4-33. Bright Field TEM Micrographs Showing Orientation Relationship Between
B2 Matrix and A2 (BCC) Alpha-Chromium Precipitates in HAM Cast Ni-50A1-5Cr
Deformed 0.35 Percent at 300K




















































Figure 4-34. Bright Field TEM Micrographs of AM Cast Ni-48.5A1-5.2Cr Deformed 0.2
Percent at 300K. Note alpha-chromium dendrites and Cr23C6 precipitates.






















































Figure 4-35. Bright Field TEM Micrographs of Slip Bands Containing <100> Dislocations
in AM Cast Ni-48.5A1-5.2 Chromium Deformed 0.2 Percent at 300K


(pw,






















































Figure 4-36. Bright Field TEM Micrograph of HAM Cast Ni-48.5A1-5.2Cr Deformed 0.3
Percent at 300K. Note the wide distribution of precipitate diameters.





















































Figure 4-37. Bright Field TEM Micrographs of HAM Cast Ni-48.5A1-5.2Cr Deformed
0.8 Percent at 300K (a) Homogeneously Distributed <100> Dislocations; (b) Precipitation
































Figure 4-38. Bright Field TEM Micrograph of As-Extruded XVIM Ni-50A1


Figure 4-39. Bright Field TEM Micrograph of XVIM Ni-50A1 Deformed 0.6 Percent at
300K
































Figure 4-40. Bright Field TEM Micrograph of XVIM Ni-49.5Al-1Cr Deformed 0.6
Percent at 300K. Slip bands consist of <100> dislocations.


Figure 4-41. Bright Field TEM Micrograph of XVIM Ni-49.5A1- Cr Deformed 0.6
Percent at 300K. Note curvature of slip band at intersection.
































Figure 4-42. Bright Field TEM Micrograph of As-Extruded XVIM Ni-49A1-2Cr


Figure 4-43. Bright Field TEM Micrograph of As-Extruded XVIM Ni-48A1-2Cr.
































Figure 4-44. Bright Field TEM Micrograph of XVIM Ni-48A1-2Cr Deformed 0.7 Percent
at 300K


Figure 4-45. Bright Field TEM Micrograph of XVIM Ni-49A1-2Cr Deformed 0.1 Percent
at 300K Showing the Intersection of Two Slip Bands of <100> Dislocations