Reactive processing and characterization of nickel aluminide-alumina functionally gradient composites


Material Information

Reactive processing and characterization of nickel aluminide-alumina functionally gradient composites
Physical Description:
vi, 204 leaves : ill. ; 29 cm.
Zhu, Hexiang, 1966-
Publication Date:


Subjects / Keywords:
Materials science and engineering thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Materials science and engineering -- UF   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1999.
Includes bibliographical references (leaves 195-203).
Statement of Responsibility:
by Hexiang Zhu.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 030368448
oclc - 41900973
System ID:

This item is only available as the following downloads:

Full Text








The completion of this work could not have been accomplished without the help,

support and encouragement of many individuals. First, I would like to express my

sincerest gratitude to Dr. Reza Abbaschian, my supervisory committee chairman, for his

invaluable guidance, support and encouragement throughout this work, and for his

numerous suggestions and corrections during the preparation of this dissertation. I would

also like to thank the rest of my committee members, Dr. R. T. DeHoff, Dr. M. J.

Kaufman, Dr. J. J. Mecholsky and Dr. B. V. Sankar, for their encouragement and advice

throughout the research of this project. The author would also like to acknowledge the

support of this research from ONR through grant N000014-94-1-0285.

I would also like to extend my thanks to my colleagues, Dr. D. Padmavardhani

Dr. D.-I. Kim, Mr. Jeffrey Mileham and Mr. F. Chen for our fruitful discussions on

various aspects of this research. Special thanks must also be given to Mr. Wayne Acree

for hours of help on electron microscopy analyses. Thanks are also due to Dr. F.

Ebrahimi and Mr. J. Bourne for helping me to start the bending tests, and Ms. X. Qian for

helping me to finish the DTA analysis.

Finally, I am deeply indebted to my parents for their love and encouragement

throughout my entire life, and my son for his patience and understanding. Lastly, I could

not have survived without the inspiration and selfless support of my loving wife, Wen

Tan, who helped me scan photos and type this manuscript.



ACKNOWLEDGEMENTS .................................................................. ii

A B STRA CT .................................. ................... .................... v


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

2. BACKGROUND/LITERATURE REVIEW ......................................... 6

2.1 Physical and Mechanical Properties of NiAl .............................. 6
2.1.1 Crystal Structure and Stability ....................................... 6
2.1.2 Thermal Conductivity and Oxidation Resistance ................. 8
2.1.3 Slip and Yield Behavior ......... ...... ......... .................... 9
2.1.4 Fracture Behavior and Fracture Toughness ...................... 9
2.2 The Reaction Process in Ni-Al System ....................................... 12
2.3 Intermetallic Matrix Composites (IMCs) ................................... 20
2.3.1 Fabrication Techniques ............................................. 21
2.3.2 Toughening Mechanism in Brittle Matrix Composites ........... 25
2.3.3 NiAl-Al203 Composites ............................................. 32
2.4 Functionally Gradient Materials ............................................... 35
2.4.1 Design and Optimization of FGMs ............................... 36
2.4.2 Fabrication Techniques .............................................. 38
2.4.3 Characterization of FGMs ........................................ 43

3.EXPERIMENTAL PROCEDURE ................................................. 46

3.1 Raw M materials ... ........ ......................................................... 46
3.2 Oxidation of Al and Ni Powders ............................................ 49
3.3 Fabrication of the Composites ............................................... 51
3.4 Differential Thermal Analysis (DTA) ....................................... 54
3.5 Analysis Techniques ........................................................ 55
3.6 Microhardness Measurement .................................... ........... 56
3.7 Fracture Toughness Measurement .......................................... 56

4. RESULTS AND DISCUSSION: REACTION PROCESS ........................... 60

4.1 Thermodynamic Considerations ......................................... 60
4.2 Reaction Process in Ni-Al and Ni-Al* Systems ............................. 68
4.2.1 Differential Thermal Analysis ...................................... 68
4.2.2 X-ray Diffraction Analysis ............................................... 76
4.2.3 Microstructural Development ...................................... 78
4.2.4 Discussion-Reaction Mechanism .................................. 84
4.3 Modeling of Reaction Kinetics of Ni-Al System ........................... 99
4.3.1 The Model ........................................................... 100
4.3.2 Results and Discussion .............................................. 106
4.4 Reaction Process in NiO-Al System ........................................ 114
4.4.1 Differential Thermal Analysis ...................................... 114
4.4.2 X-ray Diffraction Analysis ....................................... 115
4.4.3 Microstructural Development ...................................... 116
4.4.4 Discussion ........................................................... 120

5. NiAl-A1203 COMPOSITES ...................... ......... ......................... 134

5.1 Design and Processing ........................................................ 134
5.2 Microstructures ................................................................... 135
5.3 Mechanical Properties .............................................................. 144
5.3.1 Microhardness ........................................................ 144
5.3.2 Fracture Toughness ...................................................... 145
5.4 Fracture Surfaces ................................................................... 147
5.5 Discussion: Toughening Mechanisms ........................................ 158

6. NiAl-A1203 FUNCTIONALLY GRADIENT COMPOSITES ...................... 164

6.1 Design and Processing of NiAl-Al203 FGCs ............................... 164
6.2 Bulk NiAl-A1203 FGCs ........................................................... 167
6.2.1 Microstructures ................. ................................... 167
6.2.2 Microhardness ........................................................ 175
6.2.3 Fracture Toughness ................................................... 176
6.2.4 Fracture Surfaces ....................................... 177
6.2.5 D discussion .............................................................. 178
6.3 NiAl-A1203 FGC Coating ........................................................ 183
6.3.1 Microstructures ....................................................... 183
6.3.2 Microhardness ........................................................ 185
6.4 Summary ..................................... ........ ...... .................. .. 185

7.SUMMARY AND CONCLUSIONS .................................................... 191

REFFRENCES .......... ......... ... ........... ....... ..... .. ............. ... 195

BIOGRAPHICAL SKETCH ........................................................ ...... 204

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



Hexiang Zhu

May, 1999

Chairman: Dr. Reza Abbaschian
Major Department: Materials Science and Engineering

The reactive hot compaction (RHC) technique was successfully utilized to

produce bulk in-situ nickel aluminide (NiAl)-alumina (A1203) functionally gradient

composites (FGCs) as well as thin FGC coating. The FGCs consisted of four or five

layers with alumina content increasing from less than 3vol.% to about 35vol.%. The

composites were fabricated via reactive processing of the powder mixtures of nickel,

aluminum, partially oxidized aluminum (Al*) and/or nickel oxide (NiO). The technique

resulted in a gradual transition of the microstructure and properties along the thickness of

the FGC, and led to reduced residual stresses and a strong bonding between the NiAl

substrate and the FGC coating. The FGC also had higher fracture toughness than the

corresponding composites.

The phase and microstructural development for the three powder mixture systems

(Ni-Al, Ni-Al*, and NiO-Al), which occurred during reactive processing of the

composites, were systematically studied. The reaction process of Ni + Al powder

mixtures was found to be strongly affected by pressure, heating rates, heat loss and

diffusion barrier. It was found that the formation of NiAl occurred rapidly via combustion

reaction at high heating rates and with small heat loss. At slow heating rates, however,

the reaction process was slow and controlled by solid-state diffusion. The phase

formation sequence for the slow solid-state reaction was determined to be: NiAI3 -+

Ni2Al3 -+ NiAl (Ni3Al) -> NiAl.

An A1203 particle network was produced during RHC ofNi + Al* powders, while

an interpenetrating A1203 skeleton formed for NiO + Al powders. The formation of A1203

phases during RHC of NiO + Al powders was a three-stage process, with the A1203

phases coming from both the liquid and the solid state reactions. The solid state

displacement reaction between NiO and nickel-aluminides (NiAl3, Ni2A13, NiAl) is

believed to lead to the formation of an interpenetrating A1203 network in the final

product. The in-situ composites generally had higher fracture toughness than the

composites in which A1203 particles were artificially added. The highest fracture

toughness of the composites was found to be 16 MPa m12.


Intermetallic compounds are considered to be potential candidates for the next

generation of high temperature structural applications operating at temperature higher

than Ni-based superalloys [1]. Among intermetallics, the ordered compound NiAl and its

alloys are thought to be the most promising. The attractive attributes of NiAl include its

high melting temperature (1911K), good high temperature oxidation resistance, relatively

low density (5.87g/cm3), high thermal conductivity, metal-like properties above a modest

ductile-to-brittle transition temperature (DBTT), and low raw material cost. The low

density of NiAI provides one of the major advantages of NiAl for aircraft engine

applications. The decreased density results in lower self-induced stresses in rotating

turbine airfoils, and down-sized turbine disks reflecting the lower operating stresses

imposed by the reduced mass of the blades. In addition, NiAI has the high thermal

conductivity (four to eight times that of Ni-based superalloys), which provides an

improved cooling efficiency and a significant reduction in airfoil temperature and thermal

gradients. These attractive properties make NiAl a potential candidate for high

temperature structural applications [2-7]. However, like many other intermetallics, the

commercialization of NiAI as high temperature structural materials has been inhibited

due to a very low damage tolerance at room temperature, exhibited by typical fracture

toughness values around 4-6MPa mr2, and inadequate strength and creep resistance at

elevated temperature [2-5].

The use of various reinforcement materials to form composites is generally

believed to be an acceptable route to toughen intermetallics. Unfortunately, however, an

"ideal" reinforcement for many intermetallcs, such as NiAl or MoSi2 has not been

identified. In general, the reinforcements suffer from a lack of thermodynamic

compatibility and thermal expansion mismatch with the matrix. The former causes

interaction between the composites constituents during processing or at the service

temperatures, whereas, the latter creates unacceptable residual stresses during thermal

cycling. The NiAl/A1203 system, on the other hand, has been found to have good

chemical compatibility at hot pressing temperature [8-9]. In addition, NiAl/A1203

composites have been shown to have better mechanical properties than NiAl [10]. In

addition, the fracture toughness for NiAl/A1203 composites in which A1203 formed in situ

was found to be higher than that for the "synthetic" NiAl/A1203 composites [11-15]. The

in situ composites have been fabricated by the reactive hot compaction of the constituent

powders. The interpenetrating network of A1203 was believed to be the major reason for

the higher fracture toughness of in situ NiAl/A1203 composites [14-16]. However, the

mechanisms of phase and microstructure development during reactive processing of

NiAl/A1203 composites have not been fully investigated. The microstructure and phase

evolution are influenced by the processing temperature, heating rate, pressure and


It is well known that abrupt transitions in materials composition and properties

within a component, for example, a thermal barrier component involving direct bonding

of metals and ceramics, often result in sharp local concentrations of stress due to the CTE

mismatch [17-19]. The thermal stress can cause crack formation, debonding at the

interface and often the spallation of the ceramic over layer. This problem may be

overcome if the sharp interfaces are substituted by a gradual transition from one material

to the other. This approach has lead to the development of Functionally Gradient

Materials (FGMs) [17-22]. These are materials with a variation of composition and

microstructure along their thickness. The spatial variation in composition is introduced

during the fabrication to achieve desired gradient in material properties. The development

of functionally gradient materials has been shown to provide a number of distinct

advantages over conventionally bonded materials, which include reduction and optimal

distribution of stresses, improvement of thermal shock resistance and thermal fatigue

behavior, enhanced interfacial bond strength and increase of fracture toughness [17-22].

Several efforts have been made to increase the heat resistance and oxidation

resistance of Ni-based superalloy and NiAl by applying A1203 coatings [23]. However,

differences in the thermal and mechanical properties between NiAl and A1203 as shown

in Table 1.1 have made it difficult to get a well-adhered coating. Differences in the

thermal expansion coefficients give rise to residual stresses which are responsible for

debonding. NiAl/A1203 functionally gradient composites (FGCs) can serve to lessen the

effects of thermal cycling and mechanical loading by replacing the sharp discontinuity at

the interface with a gradual or stepwise change from NiAl to A1203. The primary benefit

derived from this arrangement is an increase in the tolerance of the composites to thermal

changes and mechanical damage [19, 21, 22].

Table 1.1 The Physical and Mechanical Properties of NiAl and A1203 [2-5, 24]

Properties NiAl A1203
Crystal structure Ordered cubic B2(CsCI) HCP
Melting point (K) 1911 2323
Lattice constant (nm) 0.2887 a=0.476 nm, c=1.299 nm
Density (g/cm3) 5.9 4.0
Bonding metallic and covalent ionic
Thermal expansion 15.1 x 106 8.0x 10-6
coefficient (K')
Thermal conductivity 70-80 35
Young's modulus (GPa) 199.8 (hot pressed) 387
Poisson's Ratio 0.307 0.27
Slip system <001>{ 110} (0001) <2110> basal slip

{1210 }<1010> prism plane
{1011 }<1101> ptramidal slip
Yield stress (MPa) 200-500 polycrystallinee)
200-1400 (single crystal)
Strength (MPa) 400 457 (4-point bending)
Ductility Max 4% ---
BDTT 500-700K --
Fracture toughness 4-6 polycrystallinee) 3-5 polycrystallinee)
(MPa m 1/2) 6-10 (single crystal)
Activation energy for 176-354 479
self-diffusion (KJ/mol)

The goals of this research are to determine the reaction mechanism of in situ

NiAI-A1203 composites and to develop in situ NiAl-A1203 composites and functionally

gradient composites for high-temperature structural applications. The work presented in

this dissertation consists of four major elements:

(1) study of the phase and microstructure development during reactive hot

compaction (RHC) of Ni + Al, Ni + partial oxidized Al and NiO + Al powder mixture

compacts, and determination of the reaction mechanisms during RHC;

(2) development of models for the phase evolution during RHC and guidelines for

processing of in situ composites;

(3) investigation of the effects of various parameters, such as volume fraction,

morphology of A1203 and matrix/reinforcement interface on the toughening;

(4) establishment of processing techniques of producing NiAl/A1203 functionally

gradient composites (FGCs) and characterization of the composites.


2.1 Physical and Mechanical Properties of NiAl

2.1.1 Crystal Structure and Stability

The Ni-Al equilibrium phase diagram [25] is given in Figure 2.1. NiAl is an

ordered intermetallic compound with the simple B2 CsCl crystal structure, which consists

of two interpenetrating primitive cubic unit cells, where Al atoms occupy the cube

covers of one sublattice and Ni atoms occupy the cube covers of the second sublattice

(Figure 2.2). Ordering is maintained up to the congruent melting temperature of 1911 K

(16380C). NiAl has the highest melting temperature in the Ni-Al system and is the most

stable phase based on the large negative heat of formation of -72 kJ/Mol at the

stoichiometric composition. The B2 structure is stable for large deviations from

stoichiometry (Figure 2.1), and significant long-range order is reported for both Ni-rich

and Al-rich compositions [26-27]. Deviations from stoichiometry are accompanied by a

variable defect structure, where excess Ni atoms occupy Al sites in Ni-rich compositions,

and Al-rich compositions result by the formation of vacant Ni sites [28]. The

homogeneity range makes NiAl relatively easy to fabricate by various processes from

Bridgeman single crystal growth to self-propagating high-temperature synthesis.

Weight Percent Nickel
60 70 80

0 10 20 30 40 50 60 70 80
Al Atomic Percent Nickel

Fig. 2.1 Ni-Al equilibrium phase diagram [25].

Fig. 2.2 NiAl B2 CsCI crystal structure.

2.1.2 Thermal Conductivity and Oxidation Resistance

Thermal conductivity of NiAl is 70-80 Wm''K-' over the temperature range 20-

11000C [1]. This conductivity is four to eight times that of Ni-based superalloys. This

high conductivity helps to reduce temperature gradients in parts fabricated from NiAl

during engine operation, resulting in reduced thermal stress and thermal fatigue and

effectively lowering the peak operating temperature of the part.

NiAl alloys also have an excellent oxidation resistance and have been utilized to

produce as coatings for Ni-based superalloys [29]. A combination of easy formation and

slow growth of a continuous protective A1203 scale provides the basis for this behavior.

The parabolic growth rate constant is typically two orders of magnitude lower than that

for typical Ni-based superalloys [30-31]. Doychak et al. [32], studied the transient oxides

formation on NiAl single crystals at 8000C and 11000C. They found that oxides formed

initially by nucleation and growth along the surface, but eventually a complete healing

layer of A1203 forms beneath the existing scale. Polymorphic forms of A1203 are found to

form during the initial transient stage, but are rapidly replaced by only a-A1203. Cyclic

oxidation testing appears much more severe than isothermal oxidation, due in large part

to the significant difference in thermal expansion coefficients (CTE). CTE mismatch

induced stresses cause spelling, which occurs randomly over the surface exposing fresh

material. This leads to short cyclic life times since the Al consumption rate is inversely

proportional to the oxide thickness and spelling tends to maintain thin oxide layers. Small

additions (0.1 at. %) of Zr has been found to significantly reduce spallation and slightly

reduce A1203 growth rates for isothermal conditions [33].

2.1.3 Slip and Yield Behavior

Although several types of slip systems in NiAl single crystals, such as <001>

{110}, <001> {100} in "soft" direction and <111> {110}, <111> {112} in "hard"

direction are observed [1-3], the primary slip system in NiAl is generally reported as

<001> {110}. For polycrystalline NiAl, this results in an insufficient number (three) of

independent slip systems to allow general plastic flow, as five independent systems are

required to produce the arbitrary shape change of each grain in order to maintain

coherency with adjacent grains during deformation. In general, the flow stress of NiAl is

similar to that of BCC metals, exhibiting strong temperature dependence at low

temperature that is attributed to a large Peierls stress. At intermediate temperatures there

is a plateau where yield stress is mildly dependent on temperature followed by a rapid

drop-off at elevated temperatures. The yield stress of polycrystalline stoichiometric NiAl

is reported to be independent of temperature from room temperature to 800 K [34] and to

drop rapidly above 900 K.

2.1.4 Fracture Behavior and Fracture Toughness

The fracture mode of polycrystalline NiAl is primarily intergranular at room

temperature, changes to transgranular cleavage around 400-5000C, then to ductile rupture

above 6000C [35,36]. Transgranular fracture is favored in off-stoichiometric NiAl [37]

and in coarse grained stoichiometric NiAl at room temperature and high strain rates [2].

Intergranular failure is believed to be a direct result of having only three independent slip

system in NiAl crystal. Shape incompatibility of adjacent grains leads to microcracking at

grain boundaries which coalesce, leading to catastrophic failure. Transgranular cleavage

in off-stoichiometric NiAl, on the other hand, indicates that when the grain boundaries

are as strong as the matrix, the fracture toughness is still low. Thus, the effort to increase

room temperature ductility of polycrystalline NiAl via grain boundary strengthening

seems fruitless.

The plain strain fracture toughness for binary, single phase polycrystalline NiAl at

room temperature is generally reported to be in the range of 4-6 MPa m1/2 [1-3], and

essentially independent of grain size, stoichiometry, or processing techniques. These

values are relatively low and similar to those for polycrystalline ceramics. Therefore, it is

necessary to improve the damage tolerance of NiAl before it can be used as a structural

material. There have been numerous studies undertaken in the attempt to improve room

temperature damage tolerance of NiAl using composite approach [38-45].

Nisra et al. [38] and Johnson et al. [39-40] have successfully produced in situ

composites based on the directionally solidified eutectic systems such as NiAl-Cr, NiAl-

Mo and NiAl-V. Fracture toughness about 2- 2.5 times that of monolithic NiAl was

achieved in a directionally solidified NiAl-9Mo composite [38]. Johnson et al. [39-40]

reported the fracture toughness values of 15.5, 13.2, 21.8, 24.1 and 30.7 MPam"/2 for the

directionally solidified eutectic system NiAl-12Mo, NiAl-15Mo, NiAl-34Cr, Ni-28Cr-

6Mo and NiAl-40V, respectively. Unfortunately, they also reported an inverse

relationship between room temperature fracture toughness and creep [40]. However,

alloys based on the lamellar NiAl-(Cr,Mo) eutectic displayed a good combination of

room temperature toughness and elevated temperature strength when compared to other

NiAl-based eutectics [39]. The main toughening mechanisms in these systems are

identified as crack bridging, intrinsic toughening of the NiAl phase, crack trapping and

crack deflection.

NiAl matrix composites with brittle reinforcements and prealloyed NiAI have

resulted in only moderate increases in toughness [41-43]. For example, Rigney et al. [41]

reported up to a 50% increase in fracture toughness to 9.0 MPa m/12 for hot pressed

prealloyed NiAl reinforced with 10 volume percent TiB2 particles. Kumar, et al. [10] also

obtained an increase in fracture toughness of NiAl reinforced with A1203 whiskers, up to

9 MPa m1/2 at 15 volume percent. The toughness remained at about 9 MPa mn2 up to

25vol.%, the highest level cited in the study. Plazanet and Nardou [42] investigated the

mechanical properties of hot pressed NiAl matrix composites reinforced with SiC and

ZrO2 particles. It was found that the fracture toughness of NiAl-SiCp composites was

lower than that of NiAl due to the high thermal expansion mismatch and the weakness of

the intermetallic-ceramic interface. In contrast, for NiAl-ZrO2 composite, an

improvement in fracture toughness up to 11.3 MPa ml/2 was achieved by incorporating

15 vol.% ZrO2 particles. Fracture toughness values up to 14 MPa min2 for NiAl-ZrO2

composites have also been reported by Ramasundaram et al. [43]. The main toughening

mechanism in NiAl-ZrO2 composites was considered to be a stress-induced

transformation of metastable tetragonal zirconia particles [42,43].

Recently, several in situ NiAl-ceramic composites such as NiAl-TiC, NiAI-TiB2

and NiAl-A1203 have been produced by using reactive processing technique [11-15, 44-

47]. A 50% increase of fracture toughness for NiAl-TiC was reported by Xiang et al.

[46]. A relatively large increase of fracture toughness value has been achieved in in situ

NiAl-A1203 composites. The highest fracture toughness for the composites was reported

as 16 MPa m'2 [11-15]. The toughening mechanism in the composites was thought to be

the effective crack blunting resulting from the dislocation-induced plasticity in the NiAl

grains as a result of CTE mismatch between NiAl and A1203 during cooling from the

fabrication temperature [15].

2.2 The Reaction Process in Ni-Al system

The mechanism of phase formation in Ni-Al binary system is of both fundamental

and practical interest. Reactions in this system provide a basis for reactive synthesis of

NiAl and Ni3AI intermetallic compounds. The phase diagram (Fig. 2.1) for the Ni-Al

system exhibits two solid solutions (0.1 at.% solubility of Ni in Al, and approximately 7

at.% solubility of Al in Ni at 4000C), and five stable intermetallic compounds, NiA13,

Ni2A13, NiAl, NisAl3 and Ni3Al. The reactions and phase formation sequences at Ni/Al

interfaces of Ni-Al thin films have been the subjects of a number of papers [48-52], but

with somewhat contradictory results. There is a general agreement that NiAl3 is the first

phase to form between 250 and 3200C. In addition, phase formation is believed to be

governed more by kinetics of the dominant moving species rather than by the

thermodynamic driving forces. For example, at 3500C, the thermodynamically favored

phases for the Ni-Al system are NiAl or Ni3Al, but NiAl3 and Ni2AI3 have been observed

to form [50]. Parabolic growth of NiAl3 indicated diffusion-limited kinetics with Al as

the fastest diffusing species [48-50]. However, in some work a smooth reaction front was

observed (usually with diffusion-limited growth and well-defined kinetics), while in

others an irregular reaction front was observed (proceeding preferentially along the grain

boundary). It has been suggested that impurities, especially oxygen, could be the primary

reason for the differing results.

The detailed study of the influence of oxygen impurities on the reaction of Al-Ni

couples by Heald and Barrera [51] indicates that oxygen strongly influences the course of

the reaction with an effect which depends on its location. They found that oxygen

impurities within the Al layer inhibited the grain boundary reaction channel and to some

extent inhibited the initial reaction (forming trace NiAl3 at room temperature), but did not

effect the bulk reaction (staring at 2500C). Interfacial oxygen strongly inhibited both the

initial reaction and bulk reaction, possibly by the formation of strong Al-O bond which

acts as a diffusion barrier. It was believed that a few percent O impurity within the Al

film strongly suppressed the grain boundary diffusion path, leading to the growth of a

smooth NiA13 layer. The results by Ma et at. [49] also confirm that impurities have a

strong influence on the nucleation and growth of the NiAl3 phase.

Colgan et al. [50] investigated the initial phase formation in Ni/Al, Ni/NiAI/Al,

Ni3Al/Al, and NiA13/ Ni thin films between 300 and 4250C. They found that the reaction

always started at the interface in contact with the most Al-rich film and resulted in the

formation and growth of NiAl3 at around 300 oC, unless that was the most Al-rich layer.

The NiAI3 formed in a layer-by-layer manner until all the Al was consumed. Higher-

temperature (around 4000C) anneals resulted in Ni2Al3 forming at the interface of Ni and

NiAI3. The Ni2AI3 also grew in a layer-by-layer fashion until all the NiA13 was consumed.

In addition to the initial NiA13 formation, a recent study by Edelstein et al. [53,54] has

found that the formation of a metastable Ni2A19 intermediate phase in multilayers with

periodicity in the range of approximately 40-320nm a result which is not well


There also have been a number of studies of Ni-Al bulk diffusion couples [55-57].

In these investigations, the simultaneous growth of NiAL3 and Ni2Al3 is observed, with

Ni2A13 being the dominant growing phase. The thickening of these aluminides is found to

have a parabolic dependence on the annealing time, which is indicative of a diffusion-

controlled growth process. It was found that Al atoms were the dominant diffusing

species in NiAl3 and Ni2A13 formation (around 6000C), while Ni was dominant in NiAl

and Ni3Al (at higher temperatures) [55]. At about 6000C, the main phases forming at the

Ni/Al interfaces were NiAl3 and Ni2AI3, with only thin layers of NiAl and Ni3Al present.

The situation is more complicated in NiAl with Al being the dominant diffusing species

on the Al-rich side of the phase field, and Ni on the Ni-rich side [58].

More recently, Farber et al. [59] investigated the reaction process of Ni-Al

laminated foil structures and reported the formation of Ni5Al3 phase at 650-6750C. It was

found that at 550 and 6000C the reaction kinetics was slow so that even after very long

anneals a four-layer Ni/Ni3Al /Al-rich NiAl/Ni2A13 structure was still observed. At higher

temperatures (650-7300C) all the Ni and Ni2Al3 were consumed after reasonable

exposures, and a two-phase Ni3AI /Al-rich NiAl layered structure was formed. Longer

anneals resulted in the growth of a Ni-rich NiAl layer between the Ni3Al/Al-rich NiAl

layers. At 690 and 7300C the growth of the Ni-rich NiAl proceeded until all the Ni3Al

was consumed, yielding a two-layer Ni- and Al-rich NiAl structure. At 650-6750C,

NisAl3 phase formed at the interface of Ni3Al and the growing Ni-rich NiAl layer after

long anneals. It was also found that NisA13 phase did not grow by itself but was

accompanied and preceded by the formation of Ni-rich NiAI. NisAl3 was growing only

when Ni3AI was present and started to dissolve as soon as all the Ni3Al was consumed.

Reactive synthesis of NiAl and Ni3Al intermetallic compounds has received much

attention in recent years for obtaining dense intermetallic compounds from elemental Ni

and Al powders. The synthesis, if properly controlled, can become self-sustaining

because of the highly exothermic reaction. Such a process is called high-temperature self-

propagating synthesis (SHS). In the typical combustion synthesis reaction, the mixed

reaction powders are pressed into a pellet of a certain green density and subsequently

ignited, either locally at one point (propagating mode) or by heating the whole pellet to

the ignition temperature of the exothermic reaction (simultaneous combustion mode).

Schematic representations of a typical temperature-time plot [60] and the reaction steps

for a combustion synthesis reaction are given in Fig. 2.3 and Fig. 2.4, respectively.

Comparing with conventional processing, the advantages of combustion synthesis are

primarily: (1) the generation of high temperature which can volatize low boiling

impurities and therefore, result in higher purity products; (2) the simple exothermic

nature of the SHS reaction avoids the need for expensive processing facilities and

equipment; (3) the short exothermic reaction times result in low operating and processing

costs; (4) inorganic materials can be synthesized and consolidated into a final product in

one step by utilizing the chemical energy of the reactants. NiAl and Ni3Al intermetallic

compounds [61-71], as well as other intermetallics such as MoSi2 [72,73], TiAI [74] and

NiTi [75, 76], have been successfully synthesized by several investigators using SHS.

Philpot et al. [61] studied the reaction mechanism of reactive sintering of Ni-Al

and indicated that NiAl3 and Ni2A13 were formed through solid state diffusion during

heating. The heat released from exothermic reactions of these aluminides melted

aluminum which reacted with nickel to form Ni3Al, resulting in even more heat release.

With the aluminum liquid phase and high heat, highly dense parts were produced in a

very short period of time.

The studies of pressureless reactive sintering of Ni3Al by German et al. [62,63]

indicated that the formation of a fully interconnected transient liquid phase was the key to

obtain the nearly fully dense products (> 97%). This could be achieved by properly

controlling the reaction parameters such as particle sizes, heating rate, green density,

composition, atmosphere, maximum temperature and hold time. The heating rate was

found to strongly influence the reaction process of Ni3Al. With a slow heating rate there

was more solid state reaction forming NiA13 and/or Ni2AI3 with subsequently less liquid

phase. For slow heating rates, the intermetallic compounds from solid state interdiffusion

might inhibit liquid formation at the reaction temperature. It was believed that, generally,

higher heating rates were beneficial when sintering involved a transient liquid. However,

there was a limit to the benefit of rapid heating rates, as very rapid heating rates resulted

in a loss of the process control. Therefore, intermediate heating rates at about 300C/min

proved to be most reasonable. It was also found that the Al particle size affected the

density of the final products. A particle size near 15ulm gave the lowest amount of

porosity. Smaller particles resulted in more rapid reaction (more interfacial area) and less

densification, while coarse powders led to a poor distribution of liquid in the

microstructure and subsequently less densification. The optimal Al particle size at

aboutl5um was found to form an interconnected aluminum network during processing.

& Reacants


Fig. 2.3 Schematic representation of the


'*/.: *' .1 I
01 -

(a) (b)


temperature-time curve during SHS reaction


Fig. 2.4 Typical steps in self-propagating high-temperature synthesis process (SHS) are
(a) blend powders; (b) cold compaction to controlled green density; (c) ignition; (d)
combustion wave propagates through the body; (e) final product.

Nishimura and Liu [65] investigated the reaction sintering process of Ni3Al via

vacuum hot press. It was shown that both the compressive stress and heat flow strongly

affected the reaction process and, hence, the structure and density of reaction-sintered

products. Without compression, reactive-sintered products had a relative density up to

98% and mainly Ni3Al. When a green compact self-sustained under a preloaded

compressive stress (50Mpa), a highly densified product (relative density as much as

99.3%) was obtained. It was thought that the formation of a high-temperature liquid

associated with a self-sustaining reaction is an essential factor for densification. They

described that reactive sintering of Ni3A1 consisted of three distinct reaction stages. The

initial stage involved localized self-heating of the compact by the heat released from the

reactions, such as 3A1 + Ni -+ NiAl3 and Ni + NiA13 -+ Ni2Al3. After that, the self-

propagating stage occurred, which was associated with a liquid phase with the reaction

Liquid + Ni -+ NiAl + Ni3AI. Finally, the conversion stage completed the formation of

Ni3AI through the reaction NiAl + 2Ni -- Ni3AI. It was also found that when a 50 MPa

pressure was applied immediately after the sample reached the sintering temperature

(6200C), the increased heat loss due to the improved heat conduction between the powder

compact and the alumina blocks could prevent the reaction from self-sustaining. In this

case, only NiAl3 and Ni2Al3 phases were formed.

Plazanet and Nardou [67] studied the reaction process during sintering of NiAl via

thermal explosion mode of SHS. It was found that at heating rates higher than 5C/min,

the combustion could be ignited by preheating the sample at 5380C. The combustion

could not be ignited with heating rates lower than 5C/min. In this case, the formation of

NiAl could be achieved by classical reactive sintering. The intermediate NiA13 and

Ni2Al3 phases formed during heating the Ni-Al powder compacts before the formation of

NiAl phase at 11000C. The final products in both cases for the pressureless reactive

sintering of NiAI were found to be porous in contrast to that of Ni3AI described above,

where near fully dense products (up to 98%) can be obtained. The main reason for such a

big difference is that the combustion reaction is more highly exothermic in the case of

NiAl than Ni3Al. Therefore it is necessary to utilizing external pressure to produce fully

dense NiAl during reaction process.

In order to achieve higher product densities, reactive hot compaction (RHC)

[15,68] and reactive hot isotropic press (RHIP) [69] have been developed to produce

NiAl and its composites. It was found that fully densified NiAI could be achieved if

process was appropriately controlled. A diluent or inert particle such as prealloyed NiAl

was added to the mixture of Ni and Al powder reactants to slow the combustion reaction

to a controllable level. The microstructural evolution during reactive hot compaction of

Ni + Al + prealloyed NiAl powder mixtures at a high heating rate of 1000C/min has been

investigated by Doty [68]. It was found that the reaction could not proceed to completion

when the temperature reached the ignition temperature (about 4850C) due to the

significant cooling imposed on the specimen by contact with the graphite die. The

specimen quenched after reaction had a multi-phase microstructure containing Ni,

supersaturated NiAl (possibly NiAl3 according to the microprobe analysis), Ni3Al and

prealloyed NiAI. After about 30 min. sintering at 12500C, the uniform structure with only

NiAl was achieved.

More recently, solid state reaction synthesis of NiAl from fully dense elemental

Ni and Al powder blends has been developed by Farber et al. [77,78]. The method

involves high pressure (up to 250MPa) consolidation of very fine elemental powder

blends to full density followed by controlled solid state reactive hot press at a relatively

low heating rate (20C/min). Unlike free-standing dense Ni-Al samples that reacted in an

explosive manner when heated above 4500C, no measurable self-heating of the samples

was observed in the hot press. It was found that after 10 minute exposure at 5500C, all the

elemental Al was consumed in the formation of the Al-rich NiAl3 and Ni2A13 aluminides.

Both slow heating and effective heat transfer from the sample to the massive pressure die

prevented the reaction from becoming self-sustaining and allowed the reaction to remain

solid state diffusion-controlled. After four-hour holding at 8000C under a pressure of

250MPa, a uniform NiAl microstructure with very fine grains (about 1 lm) was achieved.

2.3 Intermetallic Matrix Composites (IMCs)

Many intermetallic compounds, particularly those based on aluminum or silicon,

offer an attractive combination of low density, high melting points and good oxidation

resistance, making them attractive for high temperature structural applications. However,

most intermetallics (e.g. NiAl) exhibit a very low damage tolerance at room temperature

and inadequate creep resistance at elevated temperature. Other intermetallics that exhibit

high strengths and adequate creep resistance at elevated temperatures also have a high

brittle-to-ductile transition temperature and therefore are not damage tolerant at room

temperature. An acceptable approach to overcome these limitations of intermetallics is

the use of various reinforcement materials to form composites. The intermetallic matrix

composites can be produced by the addition of either continuous reinforcements, such as

ceramic (A1203) or refractory metal (W, Mo) fibers, or discontinuous reinforcements,

such as ceramic (SiC, TiB2) or metallic (Nb) particles, short-fibers or whiskers. Kumar

and Bao [79] and Miracle and Mendiratta [80] have recently reviewed intermetallic

matrix composites. In this section, fabrication techniques, toughening mechanisms of

IMCs, and NiAl/A1203 composites are briefly reviewed.

2.3.1 Fabrication Techniques

The high melting temperatures of intermetallics and the associated increased

chemical reactivity limit the availability of suitable container material for the processing

of intermetallic composites from the molten state. In addition, the high processing

temperatures cause extensive reaction between the matrix and the reinforcement.

Nevertheless, some investigators have achieved success in producing IMCs via melt

infiltration processes [81], electromagnetic levitation [82] and directional solidification of

eutectic alloys [38-40]. The most general success in producing intermetallic matrix

composites has come in the powder processing area. Among the powder metallurgy

routes, the simplest technique involves blending the intermetallic powder and ceramic

reinforcement and consolidating the blend to the full density. This approach is

problematic because the reinforcement distribution is limited typically by the size of

intermetallic powder particles, and the size of the reinforcement depends on what is

commercially available. More recently, due to the limited availability of pre-alloyed

powders and various processing advantages, reactive processing techniques such as self-

propagating high-temperature synthesis, thermite reaction and reactive hot compaction

have been developed to produce a wide array of IMCs.

Self-Propagating High-Temperature Synthesis (SHS). SHS has received

considerable attention as an alternative to conventional methods of producing IMCs

because of process economics and simplicity. SHS composites are formed by either

mechanically adding the reinforcement to the reactant mixture or in situ producing the

desired composites directly from the reactants. Several intermetallic matrix composites,

such as NiAl-based [68] and Ni3Al-based [83-84], have been produced by using SHS

technique. However, the products of the combustion synthesis reaction are normally

extremely porous, which is unacceptable for structural applications. Several techniques

have been investigated as a means of densifying the SHS products, such as HIPing, hot

press and use of shock waves [60].

Thermite Reaction Synthesis. Thermite reaction is an exothermic reaction which

involves a metal reacting with a metallic or non-metallic oxide to form a more stable

oxide and corresponding metal or non-metal of the reactant oxide. The process involves

the reaction sintering of raw materials, like aluminum, and metal oxides (e.g. Fe203,

Nb05s, TiO2, ZrO2, NiO, ect.) to produce in situ composites. This is a form of oxidation-

reduction reaction which can be written in a general form as

M + AO-MO + A + AH (2.1)

where M is a metal or an alloy and A is either a metal or a non-metal, MO and AO are

their corresponding oxides, and AH is the heat generated by the reaction. Because of the

large exothermic heat, a thermite reaction can generally be initiated locally and can

become self-sustaining, a feature which makes its use extremely energy efficient. The

oldest and most well-known thermite reaction, which has been used in specialized

welding and casting applications, is:

3Fe304 + 8A1 4A1203 + 9Fe ; AH = -3400kJ/mol (2.2)

Thermite reaction can be thought as a subgroup of SHS. However, many SHS

reactions start with elemental components to form the refractory compounds. Thermite

reactions have the advantages over these elemental reactions in that they may start with

naturally occurring oxides which are less expensive and more readily available than

elemental powders. Because the self-sustaining rate of thermite reactions can be adjusted

by the addition of an inert diluent, they are often used as experimental models for solid

combustion studies [85-86]. More recently, thermite reactions have become important in

the synthesis of refractory ceramic and composite materials [87]. Over the past three

decades, various investigators have used thermite-based reactions to produce composite

materials. Intermetallic matrix composites, especially aluminide-alumina composites,

such as FeAl-A1203, Ti3Al-A1203, NbA13-A1203 [88-91], have been successfully

fabricated by using thermite reaction. Like SHS, the products of the thermite reaction

also have a high level of porosity, therefore it is necessary to utilize external pressure to

produce dense IMCs.

Reactive Hot Compaction. As mentioned before, the primary disadvantage of

SHS and thermite reaction is that the products are not fully dense if the reaction rates are

not controlled appropriately. In order to achieve higher product densities, Lu and

Abbaschian [92] developed a novel technique known as Reactive Hot Compaction (RHC)

which combines SHS with the application of pressure in order to increase densification.

Reactive hot compaction (RHC) can be thought as a subgroup of powder metallurgical

processing and a variation of simultaneous SHS or combustion synthesis. RHC is a

volumetric reaction process as the consolidated powder mixture reactants are heated

uniformly within a die under pressure to the reaction temperature. Thus, the reaction

initiates at many sites throughout the powder mixture, while the application of pressure

aids densification. The reaction rate and the densification have been shown to depend on

the shape and size distribution of reactants, green density of the compacts, heating rate,

applied pressure and atmosphere [15, 62, 83]. At higher heating rate, e.g. 50 to

1000C/min, the reaction takes place by a simultaneous combustion of the powder

mixtures. It is found that to achieve full or near full densification during RHC at higher

heating rate, it is necessary to have a fully interconnected transient liquid phase for an

extended period of time [83]. To achieve this, a diluent or inert particle may be added to

the mixture of reactants to slow the combustion reaction to a controllable level. At a

lower heating rate (less than 200C/min), on the other hand, new phases form via solid

state diffusion at temperatures below the corresponding SHS ignition temperature [70].

At lower heating rate, the RHC technique is also known as solid state reactive synthesis.

Although no liquid phase forms during solid state reactive synthesis, in fine

micron/submicron size powder blends diffusion distances are short, allowing the

synthesis of intermetallics and intermetallic matrix composites to be completed at

relatively low temperatures and reasonable exposures [77, 78].

RHC has been successfully used to fabricate several intermetallic matrix

composites such as NbAl3/Nb [92], NiAl/A1203 [15,69,93], NiAl/TiB2 [69] and

MoSi2/SiC [11,12,94,95] composites. The technique has been shown to provide distinct

advantages over more traditional PM techniques. The advantages include inexpensive

starting materials, single-step consolidation, thermodynamically stable microstructures,

lower power consumption, shorter processing time and higher product purity because of

less reaction with the processing environment.

2.3.2 Toughening Mechanism in Brittle Matrix Composites

The basis for toughening brittle materials can be best presented using a fracture

mechanics approach for crack propagation within the materials. The criterion for the

propagation of a sharp crack in terms of energy balance, developed by Griffith [96], has

the form:

S F=Ey (2.3)

where E is Young's modulus, of is the failure stress, y is the fracture surface energy, a is

the sharp surface crack length, a = 7t for the plane-stress condition, and a = n (1-v2) for

the plane-strain condition and v is Poisson's ratio.

The stress state near a sharp crack tip is related to the crack length a and the

applied load o, and is described in terms of a stress intensity factor K. The relation

between these three parameters is given by [97]:

K = o*ra (2.4)

where p is a dimensionless number depending on the geometry of the cracked specimen,

and L is a unified length dimension describing the geometry of the cracked specimen.

A cracked specimen will fracture when the externally applied load increases to a

critical level (of) at which the stress intensity factor, K, reaches a critical value (Kc). The

critical stress intensity factor Kc is regarded as a material constant and is the measure of

the fracture toughness. The fracture toughness of a material is directly related to the

fracture surface energy. From Eqs. (2.3) and (2.4), it is clear that fracture toughness

increases with increasing fracture surface energy. Fracture toughness is also related to the

plasticity of the material. Plastic deformation at the crack tip blunts the crack, cutting off

the elastic singular peak of the stress distribution near the crack tip [97]. Therefore, much

higher external stress has to be applied to fracture the material.

From fracture mechanics view of point, two approaches can be used to

toughening brittle materials like intermetallics. The first approach is to reduce the

existing critical flaw size, while the second is to extrinsically create microstructural

features which increase the fracture resistance of the materials [98]. In practice, the first

approach is limited because even if inclusions and machine flaws are eliminated, many

flaws can be introduced on the surface in service. The second approach can be achieved

by microalloying the intermetallics to improve their plasticity or compositing approaches.

The main toughening mechanisms for brittle matrix composites will be briefly reviewed

in the following several paragraphs.

Crack Deflection. Crack deflection means that cracks are deflected from their

normal plane and direction of growth (tilting and twisting). The crack deflection can be

introduced by local residual stresses from thermal expansion mismatch between the

reinforcement and matrix (Fig. 2.5a), or by srtress-induced microcracking at second-

phase particles (Fig. 2.5b). The efficiency of this type of toughening depends to a large

extent on the volume fraction, shape and orientation of the toughening phase and on the

heat expansion and elastic modulus ratios of the matrix to the toughening phase [99]. In

the case of thermal residual stresses, if the difference in the coefficient of the thermal

expansion (CTE) between matrix and second phase particles is positive (i.e., Aa = acm -

ap > 0), a compressive radial stress (or) is developed at the particle-matrix boundary

along with a tensile tangential stress (ct) in the matrix; the crack is then attracted to the

particle. Conversely, when Aa < 0, Or > 0 and ot < 0, thus the crack is rejected and passes

between the particles (Fig. 2.5a). This more tortuous crack path requires additional

fracture energy and should result in greater composite toughness. Toughening by crack

deflection has been observed by several investigators in intermetallic matrix composites,

such as MoSi2/SiC [95], NiAl/TiB2 [10] and NiAl/TiC [46].

Crack Shielding. Shielding in the process zone occurs if the microstructure

elements, in the process zone surrounding the crack, display a hysteresis in their stress-

strain behavior [99]. Crack shielding can be achieved either by phase transformation in

the process zone such as the tetragonal to monoclinic phase transformation of ZrO2 in the

ZrO2 containing composite (Fig. 2.6a) or by stress-induced microcracking in the process

zone (Fig. 2.6b). The toughening mechanism in the composites containing ZrO2 is related

to a stress-induced Martensite transformation of the tetragonal ZrO2 particles to that of

the monoclinic polymorph with a higher specific volume. It is generally believed that

toughening results from the fact that a portion of energy available for fracture is

dissipated during the stress-induced transformation process. In addition, the

transformation process generates a favorable residual compressive stress environment

near the crack front as a result of the volume expansion associated with the tetragonal to

monoclinic phase transformation. Phase transformation toughening has been observed by

many investigators [97-99]. The addition of 20 vol%ZrO2 particles to NiAI results in a

higher toughness up to 14MPam"2, more than twice of pure NiAl [43].

Crack-tip stress fields and/or residual stresses can nucleate a cloud of microcracks

at weakened microstructural sites such as grain boundaries. Such a cloud can serve as to

dilate the crack-tip region and reduce the effective stress level. Opening of the

microcracks requires the contribution of energy from the applied loads. Thus additional

load must be applied to make the crack to propagate. Microcrack toughening has been

reported in several composite systems [97-99].

Crack Bridging. Crack bridging represents a major group of toughening

mechanisms in brittle materials, which refers to a phenomenon that unbroken ligaments

bridge the advancing crack surfaces behind the crack wake [100]. These ligaments could

be ductile or brittle second phase particles, whiskers, or fibers. As the crack extends,

additional energy is consumed with progressive debonding of the ligaments. These

unbroken ligaments produce traction across the crack wake, which diminish the local

crack tip stress level. When the ligaments fail in the crack wake, energy is dissipated as

acoustic waves and causes toughening (Fig. 2.7). Eventually, the bridging material fails,

either by debonding around the end or by fracture. Following failure, frictional sliding

may occur along the debonded surface. The energy dissipation upon crack propagation

thus includes terms from the energy of the debonded interfaces, the acoustic energy

dissipated upon reinforcement failure, and frictional dissipation during pullout (Fig. 2.7).

For optimal energy consumption, the fiber-matrix interface should not be too strong or

too weak. If the bond strength is high, fiber rupture will occur prior to debonding and no

bridging will develop. On the other hand, if the interface is too weak, debonding will take

place, but little energy will be consumed by the fiber-pull-out process. Crack bridging

toughening has been observed in many composite systems including both the brittle and

Residual stresses

Sm > ap

Xap > am




Fig. 2.5 Microstructural features of brittle materials toughened by crack deflection. (a)
crack deflection arising from residual stresses; (b) crack deflection resulting from

Zirconia grain

zirconia grain







Fig. 2.6 Microstructural features of brittle materials toughened by shielding. (a) crack
shielding arising from phase transformation; (b) crack shielding resulting from

Pull-out: frictional dissipation Matrix
Energy dissipated:
acoustic waves Whisker

d q

surfaces s
iIj -1 I -L-1

- Stress-free

-- L

oss of residual
strain energy


Fig. 2.7 Schematic indicating the various contributions to the steady-state toughness of
brittle matrix composites [98].

ductile reinforced intermetallic matrix composites such as NiAl/Mo, NiAl/Nb, MoSi2/SiC

and MoSi2/Nb [39, 40, 46, 95, 101-102].

Crack Blunting. Crack blunting in composite materials refers to the phenomenon

that the sharp crack tip is blunted by limited plasticity at the crack tip or when the crack

intercepts a ductile reinforcement. Accordingly, the effective stress intensity at the crack

tip is reduced resulting in higher fracture toughness.

Crack Bowing. Crack bowing is a process in which crack tends to bow between

the hard, resistant second phase particles in the path of crack propagation. The process is

analogous to dislocation being pinned by obstacles. The localized pinning of the cracks

results in an increase in the energy which is associated with a crack front since the length

of the crack front is increased due to crack bowing [103]. Therefore an increase in

fracture energy is achieved by crack bowing, resulting in toughening.

2.3.3 NiAI-AOQ3 Composites

The NiAl-Al203 system was studied by Misra who showed well established

chemical compatibility at hot pressing temperatures and limited interfacial bonding

between NiAl and A1203 [8]. Yang et al. [104] also reported that no evidence of the

reaction between the fiber and matrix was observed in the NiAI-Al203 (fiber) composite,

that interfacial bonding was too strong to promote interfacial debonding. However,

impurities in the alloy can cause fiber-matrix reaction in NiAl-A1203 composite at the

processing temperature [105]. Choo et al. [9] studied the mechanical properties of NiAl-

based composites containing a dispersion of A1N particles and A1203 fibers. The results

showed that the composite was fully dense and bonded well with the randomly

distributed A1203 fibers, and neither chemical reaction nor cracks were observed at the

fiber/ matrix interface. The creep rate of NiAl-5%AIN-30%Al203 composite was more

than five orders of magnitude slower than XD processed NiAl and about two orders of

magnitude slower than XD processed NiAl-30%TiB2. Wang et al. [106] found that the

strengthening obtained in the the NiAl-AIN-A1203 composites was related to the grain

size refinement, and the effective resistance of the grain boundaries to sliding due to

pinning by reinforcements.

NiAl-A1203 composites have been successfully produced by traditional powder

metallurgy technique using prealloyed NiAl [10, 105, 107]. The temperatures required for

the consolidation of the prealloyed NiAl powders to full density are high, usually

exceeding 13000C [107]. Misra [105] reported that single crystal A1203 fibers lost nearly

half of their original strength after composite processing. In addition to fiber strength

loss, fiber fragmentation was also observed in NiAl-A1203 composites. One possible

reason for the fiber damage was thought to be due to the creation of surface flaws

resulting from chemical reaction at the processing temperature.

Fiber damage during composite processing becomes one of the major barrier to

the use of A1203 fibers as the reinforcements for the intermetallic matrix composites.

Reactive processing using elementary powders can decrease the composite processing

temperature as low as 8500C [78], making it applicable to produce NiAI-A1203

composites. Using this technique, Doty and Abbaschian [15] synthesized NiAl-A1203

composites using the in situ reactions during the reactive hot compaction. The process

involved blending of Ni and partially oxidized Al powders with 20% prealloyed NiAl

powder. The powder mixture was cold compacted and vacuum hot pressed at 12500C for

30 min at a heating rate of 1000C/min under a pressure of 50 MPa. An in situ

interconnected network of alumina in the NiAl matrix was produced. The composites

exhibited higher fracture toughness values which increased with increasing alumina

content. The highest toughness reported was 16 MPa mi2 at about 18 vol.% A1203. They

thought that the possible toughening mechanism was the effective crack blunting

resulting from the dislocation-induced plasticity in the NiAl grains as a result of the CTE

mismatch between NiAI and A1203 during cooling from the fabrication temperature. They

also found that the transgranular cleavage dominated the fracture surface of the

composites, which was possibly due to a strong interfacial bond between NiAl and A1203.

Henager et al. [11, 12] fabricated Ni-Al/A1203 composites using solid state displacement

reactions between NiAl and NiO to form in situ Al203 network along with phases like

NiAl, Ni3Al and Ni. Their composites also exhibited fracture toughness values of about

15.5 MPa m1/2. The interpenetrating network of A1203 was believed to one major reason

for higher fracture toughness of in situ NiAl/A1203 composites [11-16]. However, how

the phase and microstructure development during reactive processing of NiA1/A1203

composites has not been fully understood so far.

Obviously, in comparison to the "mechanical" composites [10], the in situ NiAI-

A1203 composites have higher fracture toughness. The main reason is most likely due to

the different interface structure or bonding in the composites. Bowman et al. [108, 109]

studied the effect of interfacial bonding strength on mechanical properties of continuous

A1203 fiber reinforced NiAl composites. The results showed that weakly-bonded fibers

increased room-temperature toughness but provided no strengthening at high

temperatures. Strongly-bonded fibers, on the other hand, increased high temperature

strength but decreased room-temperature toughness. Chen and Mecholsky [110] have

shown that interface was important for the design of ceramic/metal multilayer

composites. They found that the strength and toughness of the Ni/A1203 multilayer

composite depended on the tortuosity of the interface, and more tortuous interface

resulted in higher strength and lower toughness. These substantiate that the interface in

NiAl-A1203 composites determines the mechanical performance of the composites.

Interfacial structures in the continuous A1203 filament-reinforced NiAl composite were

also investigated by Wang et al. [111]. They reported that a random direct bond between

A1203 filaments and NiAl matrix was achieved at an atomic level at the clean interfaces.

Modulating diffraction contrast along the interface suggested that radial residual strains

within the A1203 filaments were homogeneously distributed along the interface. These

strains were believed to be related to dislocation nucleation in the NiAl, which resulted

from the relaxation of the thermally generated residual stresses.

2.4 Functionally Gradient Materials

Functionally Gradient Materials (FGMs) are materials with a variation of

composition and microstructure along their thickness. The spatial variation in

composition is introduced during the fabrication to achieve desired gradient in material

properties. A typical FGM configuration for thermal barrier coatings is illustrated in Fig.

2.8. The FGM consists of a ceramic outer surface which is exposed to the highest service

temperatures and the inner metallic surface which, because of its high thermal

conductivity, could be cooled by liquids or gases, while at the same time, provide high

mechanical strength. The composition of the material between these two surfaces is

gradually changed from one of a ceramic matrix to that of a metal, thereby,

accommodating the stresses by means of a graded change in thermal expansion

properties. The graded change in the thermal coefficients in each area of the component

may be affected by controlling composition, microstructure and porosity between the two

composition extremes. In comparison with conventionally bonded materials, the FGMs

can provide a number of distinct advantages which include reduction and optimal

distribution of stresses, improvement of thermal shock resistance and thermal fatigue

behavior, enhanced interfacial bond strength and increase of fracture toughness [17-22].

The initial attempts in the fabrication of FGMs were targeted towards the development of

thermal stress relief type of material. However, because of the unique coupling of site-

specific properties with the gradual transition of microstructure and properties, several

other applications have emerged which include gas turbine blades, chemical reaction

vessels, cutting tools, bio-implants and thermoelectric materials [112-115]. In all these

applications, the satisfactory performance of the FGMs depends on the processing

strategies adopted and the resulting microstructure.

2.4.1 Design and Optimization of FGMs

The function of FGMs lies in the relaxation of residual thermal stress by the

introduction of a compositional gradient and in the sharing of material functions between

two sides of a single material. Watanabe [112] gave an example of the thermal stresses

generated during cooling from processing temperature in the bonding of a stainless steel

and a silicon nitride with and without graded layer (Fig. 2.9). In the case of direct

bonding, large thermal stresses which exceed the intrinsic strength of silicon nitride are

generated, while by inserting a graded layer the thermal stress is remarkably reduced.

Williamson et al. [116, 117] studied residual stresses developed at graded metal-

ceramic (Ni-A1203) interfaces during cooling using elastic-plastic finite element method

numerical model. Strong geometrical influences on stresses were observed, particularly in

constrained specimens, where the linear composition gradient resulted in reducing the

axial stress near the free edge. The results showed that for both the axial and shear

component, the disk-shaped graded specimen exhibited a significant reduction in residual

stress. The peak stress value was reduced by approximately 70% and 30% for the shear

component and axial tensile stress component, respectively. In addition, they also

investigated the effects of the interlayer thickness and composition profile on strain and

stress distributions established during cooling from an assumed elevated bonding

temperature. Compared to a non-graded interface, significant reduction of stress and

plastic strain were predicted for thicker interlayers and composition profiles that avoided

large property gradient in areas exhibiting high modulus and little plasticity. According to

the modeling results, an "optimized" Ni-A1203 FGM specimen was fabricated using HIP

by Babin and Heaps [118], but the assessment of the validity and accuracy of the FEM

calculation have not been reported.

Optimum fabrication design of Ni-MgO and Ni-TiC FGMs was investigated by

Yuan et al. [119,120] based on the finite-element simulation of thermal stress produced in

the fabrication. It was found that for the Ni-TiC FGM specimen, the optimum value of p

obtained by the elastoplastic analysis, where p is the gradient composition exponent, was

smaller than that obtained by the elastic analysis. Therefore, it is believed that the

consideration of material elastoplastic behavior is of critical important for optimization of

the metal/ ceramic FGM. Mendelson et al. [121] analyzed NiCrAlY/( ZrO2-8%Y203)

FGM coatings with five gradient interlayer designs--step layer, narrow linear, wide

linear, parabolic and exponential gradients. It was found that the parabolic and

exponential gradient interlayer designs had both the lowest combined stress and stress

gradient across the coating.

2.4.2 Fabrication Techniques

The techniques to obtain a compositional gradient in FGMs are schematically

shown in Fig. 2.10. Gases, liquids and solids can be used as the starting materials. For the

first techniques, FGM samples are prepared by chemical vapor deposition (CVD)

[20,122], physical vapor deposition (PVD) [20], and plasma spraying [123-126]. The

gaseous method allows obtaining the required FGM in the form of a film or a plate

directly. Electroplating [127] are utilized to prepare FGM coating via liquid phase. Bulk

FGMs, on the other hand, are synthesized by various processing techniques, such as

centrifugal casting [127], diffusion bonding [128], sedimentation [129-132], powder

metallurgy techniques [112, 133-135], and self-propagating high-temperature synthesis

(SHS) [136-138]. Among these, the most widely used processing techniques are thermal

spray, powder metallurgy and SHS.

Thermal Spraying. In thermal spraying, feedstock material (in the form of

powder, rod, or wire) is introduced into a combustion or plasma flame. The particles melt

in transit and impinge on the substrate where they flatten, and under rapid solidification

form a deposit through successive impingement. The technique has been traditionally

employed to produce a variety of protective coatings of ceramic, metals, and polymers on

a range of substrates. Arc spray, combustion and plasma are the main techniques

comprising thermal spray. These classifications are based on the type of heat source and

the method by which feedstock is injected. Arc-spray processes use electrically

conductive wire as feedstock, while combustion method uses powder or wire. Plasma

spraying uses feedstock in the form of powders. Due to its high operating temperature

and ability to achieve high particle velocities, plasma is most suitable for the processing

of dense, high performance deposits of refractory materials. Controlled-atmosphere

plasma spraying, such as low pressure plasma spraying (LPPS) [123], has been utilized to

deposit reactive metals and intermetallics. LPPS processing is usually conducted in a

low-pressure inert gas-filled chamber and has proven to be a highly reliable method for

depositing superalloy-type coatings on turbine blades and other aircraft engine

components. NiCrAlY- PSZ, NiCr-PSZ, NiCrCoAlY-(8%Y203-ZrO2) and Ni-A1203

FGM coatings [123-126] have been successfully manufactured by means of plasma


Powder Metallurgy. Powder metallurgical (P/M) processing of FGMs provides a

wide range of compositional and microstructural control, along with shape-forming

capability. The technique generally involves the following sequential steps with a

selected material combination of metals and ceramics: stepwise or continuous stacking of

powder premixes according to the redesigned composition profile; compaction of the

stacked powder heap and sintering with or without pressure. Compared with other

techniques, the powder metallurgical method offers a wide range of material

combinations with a close control over the graded composition. Further, thicker samples

100% Ceramic 100% Metal

Fig. 2.8 A schematic representation of the configuration
material (FGM) and corresponding properties.

Center Une

Center Line





(Unit : 1 /100 MPa)

of a functionally gradient

Fig. 2.9 Contour maps of axial thermal stress in Si3N4/Steel bonding. (a) Butt jiont, (b)
Four interlayers with linear compositional gradient [112].

100% Ceramic 100% Metal

Ga Mixture



Plasma Gun

Powder Powder

Plasma Spraying
(One Gun)

Doctor Blade


4 Sintering

Thin Sheet



Plasma Gun

Plasma Spraying

I gnition

Green Compact


Gas Mixture

o a
0 0
0 0

I cv I

Suspenslon Su penson


(Spray Forming



0 0 0 0 0 Or

Powder Metallurgy

Fig. 2.10 Schematic illustration of FGM fabrication processes [112].

can be synthesized. In particular, net shape or near net shape structural materials can be

obtained, which makes the process attractive in comparison with other processing

techniques. Many FGNIs [112, 133-135] such as Mo/PSZ, W/PSZ, SiC/AIN/Mo, AIN/W,

Ni-alloy/Si3N4 and stainless steel/PSZ have been fabricated through this technique. The

major disadvantage for the traditional P/M technique is high processing temperature and

long sintering time.

Self-Propagating High-Temperature Synthesis (SHS). As described before, self-

propagating high-temperature synthesis (SHS) or combustion synthesis process is a

powder-based process in which reactants, usually elemental constituents, when ignited,

spontaneously transform to products due to the exothermic heat of formation. With some

modifications, SHS can be used to produce functionally gradient materials from the same

combination of materials. Generally, sample preparation begins by the reaction of a series

of mixture from the powders that will react to form the constituent materials of the FGM

product. Prior to the combustion step, the samples are assembled by stacking layers of

each of the mixtuers in appropriate amounts according to the desired composition

gradient. The powder mixture is then ignited, and a combustion wave is generated that

passes through the mixture--consuming the reactants and generating the product materials

as it does so. FGMs such as TiC/Ni, TiC/NiAl, CrC/Ni, TiB/Ni, ZrO2/TiAl, (Ti-Si-O)/Ti,

TiC-A1203-NiAl and Cu/TiB2 have been prepared by combustion synthesis [136-138].

However, high level of porosity can not be avoided by using this approach.

Reactive hot compaction (RHC), a variation of SHS and P/M, as discussed before,

can be used to produce denser FGMs at lower processing temperature and less processing

time than the traditional P/M technique. RHC has been used to successfully fabricate

nearly fully dense NiAI/A1203, NiCr/NiAl and NiCr/NiAl/Al203 functionally gradient

materials [139-141] in University of Florida. Henager et al. [11,12] also fabricated NiAl-

A1203 FGC consisting 7 layers of composition ranging from Ovol.%A1203 to about

45vol.% A1203 using solid state displacement reaction between NiAI and NiO to form in

situ A1203 by means of RHC. Their technique involved a processing time of up to 5 hrs at

temperature in the range of 1250 to 13500C.

2.4.3 Characterization of FGMs

M. Finot et al. [17] numerically and experimentally studied the elastoplastic

deformation characteristics of a plasma-sprayed, tri-layered composite plate subjected to

thermal cycling from 20 up to 800C. By means of a scanning laser technique, the

changes in the overall curvature of the unconstrained plate arising from the thermal

mismatch between the constituent phases were measured. The results showed that the

introduction of a graded interlayer between the Ni and A1203 layers significantly reduces

the magnitude of the thermal stresses, delays the onset of plasticity and cracking to higher

temperature, and that the maximum tensile stress occurs at the FGM-A1203 interface.

Kesler et al. [142] has proposed an experiment method to determine processing-induced

intrinsic stress, residual stress, elastic modulus and thermal expansion coefficients of

thermal-sprayed homogeneous and graded Ni-A1203 coatings. The results showed that

when Ni-A1203 composite coatings are plasma-sprayed onto a thick steel substrate,

residual stresses as large as 200 MPa were found in the coating at room temperature. For

the fully graded coatings, where the coating was sprayed in seven discrete steps, the

tensile residual stresses in the coating were largest at the Ni-rich and Al203-rich ends, and

decreased in the central portion of the coatings. The maximum tensile residual stress of

nearly 200 MPa was found in the A1203 surface layer. The values of the in-plane Young's

modulus of the graded coatings have been measured to be as low as 54 GPa.

The thermal properties of SiC/C FGM and SiC non-FGM were evaluated by Hirai

et al. [20]. Some cracks were observed at the boundary between SiC films and the

substrate for SiC non-FGM after 40 repeated heating cycles placing the surface at 1700-

1150 K and the bottom surface at 1200-900K in the vacuum. These cracks were thought

to be the result of thermal fatigue due to the repeated heating. In contrast, SiC/C FGM

did not suffer such cracking under the same situation. A rapid decrease of the effective

thermal conductivity with the cycle time for the SiC non-FGM was observed while no

change was observed for the SiC/C FGM. It was also found the thermal shock resistance

of SiC/C FGM sample was superior to that of SiC non-FGM.

Cherradi et al. [143] investigated the thermal fatigue behavior of PSZ-(Cr-Ni

alloy) functionally gradient materials with three different concentration profiles prepared

by a centrifugal P/M process The results showed that not only FGMs had better thermal

fatigue resistance than classical coatings, but also different profiles retarded thermal

fatigue damage differently. Jung et al. [144] studied the residual stress and thermal

properties of zirconia/metal FGMs. The thermal diffusivity and conductivity of the FGMs

and directly jointed materials (DJMs) were measured by the laser flash technique. The

results showed that the stress concentration and the residual stress induced on cooling

from the sintering temperature was relaxed for the FGMs. It was thought that the TZP/Ni

FGM was excellent thermal barrier material and more stable than either TZP/SUS304

FGM or DJMs. Thermal cycling tests, conducted between room temperature and 1423K

for the NiAl-A1203 FGCs by Miller and Lannutti et al. [130,131], showed that the FGCs

are capable of withstanding extended thermal cycling and considerable high temperature

strains without substantial damage to the ceramic layer.

There are only limited research related to mechanical properties of FGMs. Miller

and Lannutti et al. [130,131] studied fracture strength and fracture toughness of a four-

layer NiAl-A1203 FGC consisting of NiAl, NiAl-30vol%A1203, NiAl-70Vol%A1203 and

A1203 fabricated by a modified sedimentation process. It was found that the values of

fracture toughness of the FGC were dependent on the location of the notch. The highest

fracture toughness reached 11 MPa mi2 when the notch was machined in the middle of

the A1203-NiAl layer, which was about twice that of the unreinforced NiAI. The bend

strength of the composites were about 3-4 times of that of both the unreinforced NiAl and

sapphire-fiber reinforced NiAl (570 MPa vs approximately 150 MPa). Gomez [145]

preliminarily investigated the fracture behavior of a five-layer NiAl-A1203 FGC prepared

via reactive hot compaction and reported that the fracture toughness values are in the

range between 13.3 to 15.2 MPa mi/2 and are dependent on the region in which the crack

nucleates. The highest fracture toughness in the FGC was found when the tip was located

in layer II, which corresponded to the toughest composite.


The reactive hot compaction (RHC) process, a combination of reactive sintering

and hot pressing, has been utilized to produce in-situ NiAl-A1203 composites and

functionally gradient composites. The process begins with an oxidation pretreatment of

elemental aluminum or nickel powders. The powders are then blended in the

stoichiometric ratio, and placed in a boron nitride coated, graphite foil lined graphite die

with a diameter of 38.4mm, and cold compacted at 50 MPa. Finally, the compacts are

vacuum hot pressed at 12000C under a pressure of 50 MPa for two hours at a heating rate

of 10C/min. The detailed experimental procedures are described in the following.

3.1 Raw Materials

The starting materials for all the composites consist of elemental Ni, elemental Al,

NiO, prealloyed NiAl and A1203 powders. The characteristics of these powders are

summarized in Table 3.1, and their typical microstructures are shown in Figs. 3.1-3.4.

Table 3.1 Characteristics of the Powders

Powder Composition Average particle Sizes Manufacturer
Ni >99.9 %Ni, <0.2% C 2.2-3.0 rm Alfa
Al 99.8 %Al, 0.1%Fe 3.0-4.5 p.m Valiment
NiO 99.0% NiO -325 mesh Alfa
NiAl 50 0.5 at.%Al 12 pm Xform
A1203 99.99% a-Al203 1.0 p.m Alfa

Fig. 3.1 Elemental Al powders, as received.

Fig. 3.2 Elemental Ni powders, as received.

Fig. 3.3 NiO powders, as received.

Fig. 3.4 Prealloyed NiAl powders, as received.

3.2 Oxidation of Al and Ni Powders

The Al powders and Ni powders were subjected to oxidation pretreatment in

moist air and normal air at different temperatures for different time to form oxide shells

of Al203 and NiO on the Al and Ni powders, respectively. As detailed elsewhere [15], the

oxidation pretreatment leads to the formation of in-situ alumina reinforcement during the

combustion synthesis. Moreover, it is shown that the presence of the moisture during the

oxidation process is necessary for improving the interconnectivity of the in-situ alumina.

Oxidation was carried out in a tube furnace. First, the powders were put into a

long quartz tube, and then the quartz tube was inserted into a preheated tube furnace. For

Ni powders, both ends of the tube were open to the air. For the Al powder, however,

compressed air was slowly bubbled (60-100 bubbles/min.) through water in a filtering

flask and subsequently introduced into one end of the quartz tube. The flask was heated

on a hot plate to raise the moisture content of the furnace atmosphere to higher levels.

The powders before and after oxidation were weighed by a digital balance to determine

the fraction of the oxide shells in the oxidized powders. Table 3.2 gives the typical

oxidation temperature and time for Al and Ni powders. By controlling the oxidation

temperature and time, different oxide contents in Ni and Al powders were obtained to

produce composites with different volume fractions of A1203. For example, after Al

powders are oxidized at 6000C for 150 hours, a 0.075 molar fraction of A1203 in the

oxidized powders was achieved and a thin alumina shell with an average thickness of

0.24[tm formed on the surface of Al powders. The oxide shells on the surface of Al

powders after oxidation was determined to be y-A1203 based on XRD analysis (Fig. 3.5).

Table 3.2 Typical oxidation temperature and time for Al and Ni powders

Powders Temperature (OC) Time (hour) Average molar fraction of oxides
in the partially oxidized powders
570 100 0.03
Al 600 100 0.06
600 150 0.075
600 300 0.085
500 0.3 0.25
Ni 520 0.3 0.47
520 1.0 0.70

10 20 30 40 50 60 70 80 90

2 Theta

Fig. 3.5 X-ray diffraction patterns of partially oxidized Al powders showing the
formation of y-Al203 shells on the surface of Al powders.

3.3 Fabrication of the Composites

Fig. 3.6 shows the procedure for the synthesis of the NiAl-A1203 composites and

functionally gradient composites (FGCs) using reactive hot compaction (RHC) technique.

All the powder mixtures with determined compositions were placed in a plastic bottle

with 5-15 steel balls, each 3/8" in diameter, and mixed in a rotary mixture for at least 10

hours. After mixing, the steel balls were removed and the powder was sieved through a

#100 screen (150/pm opening) to remove any large agglomerations. When a uniform

composite was fabricated, the weighed powder mixtures were placed in a boron nitride

coated, graphite foil lined graphite die with a diameter of 38.4mm, gently tapped and cold

compacted at 50 MPa. When a multi-layer functionally gradient composite was

fabricated, the powder mixtures corresponding to different layers were placed in the

graphite die, gently tapped and stacked layer-by-layer, and finally cold compacted at 50

MPa (Fig. 3.6). After cold compacting, the powder mixtures had a green density of 60-

65% of theoretical. The compacts were then vacuum hot pressed at 12000C and at a

pressure of 50 MPa for two hours at a heating rate of 10C/min (Fig. 3.7). The hot press

ram was controlled by manual hydraulic pump fitted with a pressure gage. The

temperature was monitored via optical pyrometer sighted through a quartz access window

in the hot press chamber. Several tests were run utilizing type K thermocouples inserted

into the graphite die and the samples to obtain more precise temperature measurement at

low temperatures and for optical pyrometer calibration.

Starting elemental powders

Partial oxidation of Al and/or Ni

Powder mixing in the stoichiometric
ratio with addition of NiAl diluent

Powder blending in a rotary mixer
for at least 10 hours

Stepwise stacking of layers

Cold compaction at 50 MPa

Hot press at 50 MPa, 12000C in a
BN-lined graphite die

Fig. 3.6 Flow sheet showing the procedure for the synthesis of NiAl-Al203 Functionally
Gradient Composites (FGCs) using Reactive Hot Compaction (RHC).

1400-- 50
1400 T=12000C
1200 -

0 1000-

S800 -

E 600-

200 Furnace Cooling

0 50 100 150 200 250 300 350 400

Time (minute)

Fig. 3.7 Reactive Hot Compaction (RHC) processing cycle for NiAl-A1203 composites.

The composite samples with two different sizes were prepared. The samples for

fracture toughness measurement had a thickness of about 6 mm, while that for the study

of the reaction process was thinner, with a thickness about 2 mm. During RHC, the

thinner samples were quenched at temperatures ranging from room temperature to

1200C by turning off the power immediately after the temperature reached the

determined value. Two kinds of NiAl-A1203 functionally gradient composites were also

fabricated, one bulk with a thickness of about 6 mm, and the other thin coating on NiAl

substrate with a thickness of about 900 p.m.

3.4 Differential Thermal Analysis (DTA)

The differential thermal analysis (DTA) of the powder mixtures, cold compacted

at 50 MPa, was carried out in a DTA to obtain information about reaction temperatures

during RHC. As the reactive synthesis techniques are sensitive to the heating rates

adapted, the effect of heating rate on the formation of the various phases was studied at

four different heating rates, 5, 10, 20 and 500C/min. The argon flowed at 100 ml/min.

during DTA analysis.

The temperature changes of the samples during RHC were also measured by a K

type (NiCr-NiAl) thermocouple, which was inserted near the center of the samples. An

A1203 powder compact was used as the standard for the measurement. The temperature

difference (AT= Ti T) between the sample (TI) and the standard (T) was also measured

during RHC.

3.5 Analysis Techniques

All metallographic specimens were cut from the hot press disks using an electric

discharge machine (EDM) and mounted in bakelite. After grinding to 600 grit on wet SiC

paper, each sample was diamond polished with 6 jim then 1 pun diamond paste.

Metallographic specimens were subsequently analyzed in the as-polished condition, or

etched with Kroll's Reagent to reveal the grain boundaries or deep etched with saturated

molybolic acid for 20-25 minutes to reveal the alumina network. The compositions of the

etchants are given in Table 3.2.

Table 3.2 Etchants for Metallographic Specimens

Etch Name Kroll's Reagent Saturated Molybolic Acid
Purpose To reveal Grain Structure To dissolve NiAl matrix to
reveal Al 203 Network structure
Formula 7mlHF+14mlHNO3 +100ml 100gm Mo 03 +50ml HF +150ml
H2 0 H20
Technique Swab for 20-30 seconds, Immerse in etchant in ultrasonic
rinse with water, then cleaner for 20-25 minutes, rinse
methanol, then blow dry. with water, then methanol, then
blow dry.

Phase identification in as-synthesized and quenched samples was performed

employing X-ray diffraction (XRD). A Philips diffractometer (APD 3720) with a Cu X-

ray tube (XcuKal=0.154056nm) was employed operating at 40 kV and 20 mA. Scanning

was performed in a step mode with a 0.020 step and a 1 s exposure at point in a 10-90

diffraction angle range. Microstructures and chemical composition of the samples were

studied employing optical microscope (Nikon Epiphot), electron microprobe (Jeol

Superprobe 733), and a scanning electron microscope (SEM) JSM-6400 with an energy

dispersive spectrometer (EDS) Link AN10000. Quantitative EDS analysis was based on

K analytical lines from the spectra acquired at the 15 kV accelerating voltage. Aluminum

was used as a standard, and ZAF correction was performed using a standard procedure

from the Link software package.

3.6 Microhardness Measurement

Microhardness measurements via Vickers indentations were performed by

applying a 1 Kg load for a period of 12 second load time on the flat surface of polished

samples. The microhardness near the interfaces between different layers in functionally

gradient composites was determined by taking measurements diagonally across the

interface in 10 um intervals.

3.7 Fracture Toughness Measurement

Four-point bend test fracture specimens were cut from the as-hot pressed disks by

EDM with the dimension given in Fig. 3.8. The specimens were ground on wet SiC paper

followed by diamond polishing up to 1 p.m. Chevron notches of the specimens were

introduced by carefully making two cuts with a 0.15mm diamond watering blade using a

specially made fixture so that the notches could be cut into a desired angle and depth.

Finally, all specimens were ultrasonically cleaned in acetone for ten minutes prior to

testing. The bend tests were conducted on an Instron model 1125 testing machine at a

constant cross head speed of 8.6x104 mm/s at room temperature in air. The inner and

outer spans, 10 and 20 mm, respectively, were applied to the specimen via SiC roller

fixtures, which were free to tilt and rotate to eliminate any lateral forces during the

application of the testing load.

The peak load, Pmax from the load vs. displacement curves was utilized to estimate

the fracture toughness, Kmax, with the aid of the following equation [146]:

Kmax _P Ymin (3.1)
B W-

where B and W are the width and height of the bending bar, respectively and Y*m, is the

minimum value of the dimensionless stress intensity factor coefficient as a function of

relative crack length for the particular specimen used. Y for the chevron-notched

samples is given by [147]:

min =(3.08+ 5ao + 8.33a )( )[ + 0.007(S' )2][a-ao] (3.2)
W W 1- ao

where a0 = a_,a = a, S is the outer span (20 mm), S2 is the inner span (10 mm). ao

and al are defined in Figure 3.8. This allows the direct calculation of fracture toughness

from peak load and the specimen geometry.

For the FGCs, the location of the tip and the bottom side of the notch

(corresponding to the origin and propagation of the crack, respectively) was found to

affect the fracture toughness values because of the anisotropy of the material. Fig. 3.9

schematically shows the location of the chevron notch in the four different fracture

toughness tests performed on the FGCs. For example, in the first test, called II -+ IV, the

tip of the chevron notch is located in layer II, and the bottom side of the notch is in the

side corresponding to layer IV. In this case, the crack is expected to initiate in layer II

and propagate toward layer IV. The rest of the other tests follow the same description,

giving rise to different location of crack initiation and different path of crack propagation.









Fracture surface

Fig. 3.8 Configuration of four point bend test on chevron notched specimens (all
dimensions in millimeters).


II --> IV = Crack initiation in layer II, crack propagation toward layer IV.
III --> I = Crack initiation in layer III, crack propagation toward layer I.
I --> IV = Crack initiation in layer I, crack propagation toward layer IV.
IV --> I = Crack initiation in layer IV, crack propagation toward layer I.

Fig. 3.9 Schematic showing the crack initiation and crack propagation direction for
chevron-notched four point bend tests for NiAl-A1203 functionally gradient composites

Fracture Toughness Tests for FGCs
1 2 3 4

II-- IV III-- I I -0 IV IV --I


Reaction hot compaction (RHC), a combination of reactive sintering and hot

press, is a processing technique for simultaneous synthesis and densification of

compound materials from the elemental constituents. In this work, the RHC technique

was utilized to fabricate NiAI, NiAI-A1203 composites and NiAl-A1203 functionally

gradient composites. The powders such as Al or partially oxidized Al, Ni or partially

oxidized Ni, and NiO were mixed and cold compacted, followed by reactive hot

compaction of the powder compacts at 12000C under a pressure of 50 MPa for two hours

at a heating rate of 10C/min. In this chapter, the reaction processes for three systems

consisting of Ni-Al, Ni-partially oxidized Al, and NiO-Al are systematically studied. The

phase and microstructural development and the reaction mechanisms are discussed.

4.1 Thermodynamic Considerations

During combustion synthesis, once the reaction initiates, extremely high

temperatures can be achieved in a very short time due to the highly exothermic nature of

the reactants. It is therefore reasonable to assume that a thermally isolated exothermic

system exists because there is very little time for the heat to disperse to its surroundings.

Therefore, the maximum temperature to which the product is raised can be assumed to be

adiabatic temperature, Tad.

Consider an exothermic combustion synthesis reaction in which a green reaction

powder mix, at an initial temperature To, is ignited under adiabatic conditions at an

ignition temperature, Tig. In order for the reaction to ignite at Tig, the reactants need to be

heated from To to Tig. Therefore, the amount of heat, H(R), needed to do this is given by

the following equation [60]:

H(R) = nCp(Ri)dT + nL(Ri) (4.1)
To- Tig

where ni, Cp(Ri), and L(Ri) are the reaction stoichiometry coefficients, heat capacities,

and the phase transformation enthalpies (if the reactions undergo a phase change, e.g.,

melting), respectively, of the reactants, Ri. The value of H(R) is indicated in Fig. 4.1.

Since the combustion synthesis reaction is initiated at Tig, the heat of the reaction under

this condition is given by AH(Tig), also indicated on the H-T plot in Fig. 4.1. Since H(R)

is needed to heat the reactions from To to Tig, the amount of heat available to be absorbed

by the products under adiabatic condition is, therefore, H(P). The latter raises the

temperature of the products from Tig to Tad(To). Thus, the total heat generated by the SHS

reaction, AH(Tg), is distributed between heating up the reactants from To to Tig, and

heating the products from Tig to Tad:

AH(T,) = -[H(R) + H(P)] (4.2)

Similarly, H(P) can be represented by Eq.(4.3):

H(P) = d(TO) nIC,(P)dTr+ >nL(Pj) (4.3)

where nj, Cp(Pj), and L(Pj) are the reaction stoichiometry coefficients, heat capacities,

and the phase transformation enthalpies (if the products undergo a phase change, e.g.,

melting), respectively, of the products, Pj. Therefore, the maximum adiabatic temperature,

Tad(TO), which is indicated in Fig. 4.1, will depend on the initial temperature (i.e.

preheating). For example, increasing preheating temperature from To to T1 will decrease

H(R), increase H(P), and increase Tad(To) to Tad(TI). Increasing To to Tig will decrease

H(R) to zero and all of AH(Tg) will be available to be absorbed by the products,

resulting in an adiabatic temperature of Tad(Tig). Under this condition the reaction is

ignited under simultaneous combustion mode. This is the condition used in the present

experimental investigation.

Since the enthalpies of reactants and products are commonly given at 298K,

AH(Tg) can be calculated using the following relationship:

AH(T,g) = AH(298)+ I9 jnCp(P) nCp(R1,)dT+ InL(P,)- inL(R,) (4.4)
[298-T, 298- Tg J

where H(298) is the reaction enthalpy at 298K. Substituting Eqs. (4.1), (4.3), and (4.4)

into Eq. (4.2), and rearranging results in Eq. (4.5), provides a means by which Tad can be


AH(298)+ f, njCp(Pj)dT + InjL(Pi) = 0 (4.5)

Most combustion synthesis reactions are conducted under nonadiabatic conditions

on account of heat losses incurred from the reaction. Under these conditions, the

maximum combustion temperature, Te, will be less than Tad by the extent of the heat loss

Q, as indicated in Fig. 4.1, i.e., Tc(To), is that for SHS reaction conducted in the

propagating mode with an initial temperature of To.

Assuming reactions for the two systems, Ni + Al and NiO + Al, are adiabatic

during reactive processing, the maximum temperature Tad can be calculated using

available thermodynamic data [148]. The ignition temperature for Ni + Al during RHC

was measured to be 4800C (see Section 4.2). Fig. 4.2 shows the enthalpy-temperature

curves for reactive sintering of NiAl using Ni + Al powder mixtures. It can be seen that

the maximum temperature Tad will reach 1911K, and the product NiAl will almost

completely melt. Thus, the thermodynamic analysis indicates the possible formation of

molten NiAl during reactive sintering of Ni + Al powders. The previous results have

shown that it is necessary to add a certain amount of diluent to obtain dense product

during reactive sintering of NiAl [68]. The effect of the diluent such as prealloyed NiAl

and A1203 particles on the reaction process is shown in Fig. 4.2, which indicates that the

addition of diluent can decrease the amount of the liquid phase during reactive sintering.

In addition, during the actual processing, considerable amounts of heat released by

exothermic reaction may be transferred to the surroundings, and the maximum

temperature could be much lower. As shown later, the heat loss during RHC of NiAl is

strongly affected by the heat transfer between the sample and the surrounding graphite

mold in this investigation.

The ignition temperature for NiO + Al powder mixtures was measured to be

around 1000K [85]. Figs. 4.3 and 4.4 show the enthalpy-temperature curves for reactive

sintering of the NiAl-A1203 composites using NiO + Al powder mixtures with the

addition of NiAl and A1203 diluents. It can be seen that the maximum temperature Tad

AH(T i, )-(H(R) +H(P)] ^^

---- -------------
H(R) Heat loss A Q

H(P) .. .

To TI Ti To(To) Td (T) Ta(T,)TT (Tg)
Temperature (K)

Fig. 4.1 Schematic representation of the enthalpy-temperature plot for reactants and
products in a reaction system that involves no phase changes in reactants and products




-- -1.25NiAI
-- Ni+AI+O.1AI203
-..- NiAI+0.1AI203

' .1. I I I. .I II' 1 I I ..1 I .I I

AI203 ,mp

Al, mp
NiA, mW ,

ri 0
-4 E4
r- 0

I- Z Z
n I I I I I< :
21 '* **21 "o "



1500 2000 2500 3000

Temperature (K)

Fig. 4.2 The effect of diluents-the prealloyed NiAl or A1203 particles on the enthalpy-
temperature curves for Ni +A1 and NiAl.



















S 1 IAl2 0 m

A1203 m]

NiO mpt
NiAl mpt

Al mpt



Al bpt y=9

Io f -" y =9
Pt 0

y" yp =3

(* 1 0 o -

'U o 'o 'U 'oU '
0 -
1 E1 EI E4 ETE4I

0 500 1000 1500 2000 2500 3000 3500 4000

Temperature (K)

Fig. 4.3 The effect of the addition of prealloyed NiAl on the enthalpy-temperature curves
for reactive sintering ofNiO +Al powder mixtures.
















3000 3500 4000

Fig. 4.4 The effect of the addition of alumina particles on the enthalpy-temperature
curves for reactive sintering ofNiO +Al powder mixtures.

500 1000 1500 2000 2500

Temperature (K)

Al bpt
A 203 mpt x=0
NiO mpt
NiAl mpt xr=1
Al mpt

L -

".- x=3

Z ---x

------ -

. -----------t i
O) r M E 1 ^ !.

will reach about 3800K, and the product, NiAl and A1203, will completely melt. Because

the thermite reaction of NiO + Al is more highly exothermic compared to the reaction of

Ni + Al, it can result in a large amount of liquid phase during reaction sintering and even

some gaseous phases such as Al, which makes the reaction uncontrollable and may lead

to a high-level porosity in the final product. Accordingly, to make the reaction

controllable and obtain dense products, it is necessary to decrease the combustion

temperature by using slow heating rate to enhance the heat transfer between the sample

and the surrounding graphite mold and/or by adding the diluent such as prealloyed NiAl

and A1203 particles to absorb some heat. From Figures 4.3 and 4.4, it can be seen that the

addition of A1203 particles is more effective in reducing the combustion temperature.

4.2 Reaction Process in Ni-Al and Ni-Al* Systems

4.2.1 Differential Thermal Analysis

DTA has been performed on the green compacts ofNi + Al and Ni +Al* (partially

oxidized Al) powder mixtures using heating rates of 5, 10, 20 and 500C/min. Fig. 4.5 and

Fig. 4.6 show the DTA curves for the above two systems at three heating rates. The

curves indicate the occurrence of several exothermic and endothermic reactions at

temperatures summarized in Table 4.1. Since the starting temperatures for the exotherms

are not well defined, the temperatures corresponding to the maximum points in the curves

are reported for the occurrence of the exothermic reactions. The starting temperatures for

the endotherms, on the other hand, are better defined and are also reported in Table 4.1.

For Ni + Al powder compacts, three exothermic peaks and two endothermic peaks were

observed at lower heating rates (5, 10 and 200C/min), while only one strongly exothermic

peak was detected at higher heating rate (50C/min), starting at 5500C (Fig. 4.5). For Ni +

Al* powder compacts, two exothermic peaks and three endothermic peaks were observed

on DTA curves (Fig. 4.6). It can be seen that the temperatures for the exothermic peaks

increased with increasing heating rates. For example, in the case of Ni + Al* system, the

temperatures of occurrence of the first exotherm are 550, 585 and 6060C, and those for

the second exotherm are 973, 1007 and 10470C at heating rates of 10, 20 and 50OC/min,

respectively. However, the heating rate almost did not affect the endotherm temperatures.

Also the temperatures for the occurrence of these peaks are independent of composition.

For Ni + Al* powder compacts, the first exothermic peak is diffuse and not well-defined,

and its height is much lower than the first exothermic peak of Ni + Al samples. While the

first and third endotherms for Ni + Al* are very weak, the second one is strong. The

second exotherm for Ni + Al* system is asymmetric and seems to be shouldered

suggesting an overlap of two exotherms.

The amount of energy released in the case of the two exotherms may be compared

qualitatively by considering the areas under the peaks. As seen in Fig. 4.6, the heat

liberated during the first exothermic event is lower than the heat released during the

second one. Also the heat liberated increased with increasing heating rate. For example,

in the case of Ni + Al* powder compacts, the exothermic peaks for the first exotherm at

heating rates of 10 and 200C/min are not well defined, in contrast with that at 50OC/min.

The actual hot press conditions could not be duplicated in the DTA due to

insufficient heating power and instrument limitations to apply direct pressure to the

sensitive balance feature incorporated for TGA. To better understand the reaction process

of Ni + Al powder compacts during reactive hot compaction, the temperature-time curve

and temperature difference of the sample and the standard (A1203) during RHC were

measured by using K type thermocouple inserted inside the sample and the standard.

Three tests were conducted; the first one at a heating rate of 100C/min, the second one at

500C/min and the third one using sample well insulated with A1203 at a heating rate of

500C/min. From the temperature-time curve (Fig. 4.7) and special DTA curve (AT vs. T

curve during RHC) (Fig. 4.8), it can be seen that for the well-insulated sample the

ignition temperature is about 4800C. This ignition temperature is very close to 4850C

estimated by Doty [67] at a heating rate of 1000C/min under similar conditions. The

application of pressure in the hot press reduced the ignition temperature from 5500C to

4800C. The maximum temperature was measured as about 12000C, which is much lower

than the theoretical temperature, 16380C, predicted by the thermodynamic calculation

presented in Section 4.1.

The special DTA curves for the other two Ni + Al samples at the normal hot press

conditions at heating rates of 10 and 500C/min are shown in Fig. 4.9. Although only one

peak was observed for the sample at a heating rate of 500C/min, the temperature increase

of the sample was very limited, at a range of about 200C, in contrast to about 7000C for

the insulated sample shown in Fig. 4.8. At a heating rate of 100C/min, only one very weak

peak was observed. Unlike free-standing dense Ni-Al samples (Fig. 4.5) or the well-

insulated samples (Fig. 4.8) which reacted in an explosive manner when heated above

4800C, no measurable self-heating of the samples was observed in the hot press at a slow

heating rate of 100C/min (Fig. 4.9). As discussed later, both slow heating and effective

heat transfer from the sample to the massive graphite mold prevented the reaction from

becoming self-sustaining and allowed the reaction to remain solid state diffusion-


Table 4.1 Temperatures corresponding to the exothermic and endothermic events in the
Ni + Al and Ni + Al* powder compacts obtained from DTA at heating rates of 5, 10 and

Heating 50C/min 100C/min 200C/min 500C/min
Samples Exo- Endo- Exo- Endo- Exo- Endo- Exo- Endo-
therm therm therm therm therm therm therm therm
(C) (0C) (0C) (0C) (0C) (0C) (0C) (0C)

Ni +AI 557 572 587
642 640 639
648 649 653 657
863 873 870
892 908 937

Ni +Al* 440 550 585 606
640 640 640
664 664 665 662
867 865 867 862
939 973 1007 1047

300 -- 500C/Min

=U I
1 200 -


100 -

0 -

200 400 600 800 1000 1200
Temperature (oC)

Fig. 4.5 DTA curves of Ni + Al powder compacts at heating rates of 10, 20 and






200 400 600 800 1000 1200

Temperature (OC)

Fig. 4.6 DTA curves of Ni + Al* powder compacts at heating rates of 10, 20 and





0 200 400 600 800 1000
Time (s)

Fig. 4.7 The sample temperature vs. time curve for the alumina-insulated Ni+Al powder
compacts during RHC at a heating rate of 500C/min.

600 -

400 -

200 -

[ I
0 100

200 300 400 500 600 700
200 300 400 500 600 700

T (oC)
Fig. 4.8 The temperature difference (AT=T1-T) vs. T curve for the insulated Ni + Al
powder compacts during RHC at a heating rate of 500C/min (TI, T are the temperatures
for the sample and the A1203 standard, respectively).


-- 10C/min
20 -

15 -




-5 I I
0 200 400 600 800 1000 1200

T (oC)

Fig. 4.9 The temperature difference (AT=TI-T) vs. T curve for the Ni+Al powder
compacts during RHC at heating rates of 10 and 500C/min without insulation (normal

4.2.2 X-Ray Diffraction Analysis

The evolution of the phases during heating of the Ni + Al powder mixtures in

vacuum at 50 MPa was investigated by halting the heating at different temperatures. The

XRD patterns of the samples heated to 6000C at a heating rate of 500C/min with and

without insulation are shown in Fig. 4.10 and Fig. 4.11 respectively. It can be seen that,

for the well-insulated samples, the final products only consist of NiAl phase (Fig. 4.10),

while for the normally hot pressed samples, the final products are composed of Ni2Al3,

NiAI, Ni3Al and Ni multi-phases (Fig. 4.11). The results indicated that the heat exchange

condition controlled the completion of combustion reaction from Ni + Al to NiAl, and

that heat loss resulting from the heat exchange between the samples and graphite mold

may prevent the combustion reaction from becoming self-sustaining.

XRD diagrams for the reactive hot pressed Ni + Al powder mixtures from room

temperature to 8500C at a heating rate of 100C/min are shown in Fig. 4.12. It can be seen

that after milling, X-ray diffraction patterns do not show any peaks which are different

from those of Al and Ni. At a temperature less than 400C, no new phases were detected

either. As the temperature is raised to about 4500C, peaks related to the formation of

NiA13 appear, while the Al peaks weaken. When the temperature reaches 5000C, the Al

peaks are completely replaced by those corresponding to Ni2Al3, and the NiAl3 peaks

weaken. At 5500C, the NiAl3 peaks disappear and the peaks corresponding to NiAl and

Ni3Al appear. From 5500C to 6800C, the four phases Ni2A13, NiAl, Ni3AI and Ni co-exist,

however, NiAl and Ni3AI grow at the expense of Ni2A13 and Ni. As the temperature is

raised to 7500C, the Ni peaks disappear. When the temperature reaches 8500C, NiAl

becomes the major phase, only minor Ni3AI exists and the Ni2AI3 peaks disappear. After

about two hours at 8500C, the reaction is complete and the samples only consist of NiAl.

The peaks of NiAl in all the XRD patterns of the samples halted at different temperature

(550 8500C) were slightly shifted towards higher diffraction angles in comparison to

those of the NiAI after two-hour holding at 8500C (Fig. 4.13), indicating a smaller lattice

parameter than that of stoichiometric NiAI. NiAI has a wide homogeneity range and both

Ni-rich and Al-rich compounds have lattice parameters smaller than that of the

stoichiometric NiAl [3,4]. Therefore, it can be suggested that after heating to 8500C via

RHC, the synthesized sample consist of a mixture of the hypo- and hyperstoichiometric

NiAl. This will be confirmed by the EDS analysis presented later.

The evolution of the phases during heating the powder mixtures of Ni + Al* was

also investigated by halting the heating at different temperatures during RHC. XRD

diagrams for the samples from room temperature to the sintering temperature under 50

MPa at a heating rate of 100C/min are shown in Fig. 4.14. The phase development for the

powder mixtures of Ni + Al* is similar to that of the Ni + Al, except that the reaction

seems to be delayed somewhat. After milling, X-ray diffraction patterns do not show any

peaks which are different from those of Ni, Al and y-A2l03. At 4500C, no new phases

were detected, in contrast with that a large amount of NiAl3 was detected in the case of

non-oxidized Al + Ni powder mixtures (Fig. 4.12). The comparison indicated that the

existence of A1203 on the surface of Al delayed the formation of NiAI3. For partially

oxidized Al powders, NiAl3 formed between 4500C and 5000C (NiA13 was detected in

samples heated to 5000C, not shown in Fig. 4.13). As the temperature is raised to 5500C,

peaks related to the formation of Ni2AI3, NiAl and Ni3Al appear, while the NiAI3 peaks

disappear and the Al and Ni peaks weaken. From 550C to 6800C, the four phases Ni2Al3,

NiAl, Ni3Al and Ni co-exist, as well as y-A1203. When temperature reaches 8500C, the

NiAl becomes the major phase with only minor amounts of Ni2Al3 and Ni3Al. At 10000C,

there still exists minor Ni3Al. Further, the y-Al203 peaks are very weak. After about two

hours at 12000C, the reaction completes and the sample consists of NiAl and a-A1203.

The transformation ofy-Al203 to the more stable phase a-A1203 occurred at a temperature

between 1000 and 1200C. It has been reported by several investigators that the

transformation of y-A1203 to a-A1203 occurs between 1000 and 13000C [68, 149, 150].

4.2.3 Microstructural Development

Microstructures of the hot pressed powder compacts (Ni + Al) heated to different

temperatures from 4000C to 12000C at a heating rate of 100C/min are shown in Fig. 4.15.

Intermediate phases such as NiA13, Ni2Al3 and Ni3Al are produced during reactive hot

press depending on the temperature. At 4000C, no obvious phases other than Ni (seen as

gray in the micrograph) and Al (black) were seen or detected by EDS. The spherical Al

powers deformed due to the application of 50 MPa pressure, resulting in a continuous Al

network surrounding the Ni powders. As temperature increases, NiAl3 forms at the

interface between Ni and Al. At 4500C, a large amount of NiAl3 is produced with only a

small amount of Al remaining (Fig. 4b,c). At 5000C, Ni2Al3 grows at the expense of

NiAl3 and Ni and becomes the main phase in the sample. When the samples were heated

to 5500C, both NiAl and Ni3Al were detected at the interface between Ni and Ni2Al3. At

6800C, NiAl became the main phase in the sample, but the four-phase

Ni/Ni3A1/NiAl/Ni2Al3 structures co-existed over a large temperature range (550-7500C).

At 6800C, NiAl was determined to be Al-rich (53at.%Al 47at.%Ni) by EDS.

30 40 50 60 70

2 Theta

Fig. 4.10 XRD patterns of the final product formed when the insulated Ni + Al powder
mixtures were reactive hot pressed to 6000C at a heating rate of 500C/min.

10 20

30 40 50 60 70 80

2 Theta
Fig. 4.11 XRD patterns of the final product formed when the Ni + Al powder mixtures
were reactive hot pressed to 6000C at a heating rate of 500C/min without insulation.

80 90

10 20

10 20

30 40 50 60

70 80

2 Theta

Fig. 4.12 XRD patterns of the intermediate products formed upon heating to the
temperature indicated during reactive hot compaction of Ni + Al powder mixtures at a
heating rate of 100C/min without insulation.


8500C + 2 hours NiAI



50 60 70 80 90

2 Theta

Fig. 4.13 XRD patterns of the reactive hot compactioned Ni + Al powder compacts
heated to 8500C and held for 0 and 2 hours at a heating rate of 100C/min, showing the
NiAl peak shift.

10 20

30 40

50 60 70

2 Theta

Fig. 4.14 XRD patterns of the intermediate products formed upon heating to the
temperature indicated during reactive hot compaction of Ni + Al* powder mixtures at a
heating rate of 100C/min.

Only a two-phase NiAl/Ni3AI structure was observed when the samples were heated to

8500C (Fig. 4.15h). A Ni-rich NiAl was detected at the interface between NiAl and Ni3Al

after heating in the range of 750 and 8500C (Figs. 4.15g, h). After a two-hour anneal at

8500C (or continuing to heat to 10000C), no additional reaction was detected, and the

sample contained a uniform NiAl structure with small amount of A1203 plus some

porosity (Fig. 4.15i). Finally, after two hours at 12000C, a near fully dense NiAl was

achieved (Fig. 4.15j).

The presence of a thick oxide shell on the aluminum powders changed the

microstructural evolution during reactive processing. Fig. 4.16 shows microstructures of

the reactive hot compacted powder compacts of Ni + Al* halted at different temperature

(from 4500C to 12000C) at a heating rate of 100C/min. At 4500C, Ni (as seen as white

gray in the figure), Al (dark gray) and A1203 shells (black) on the spherical Al powder

surfaces were observed in the SEM pictures. Some regions with Al powders with A1203

coatings were observed in the microstructures, in contrast with a continuous Al network

surrounding Ni powders in the case of Ni + Al powder compacts. As temperature

increased, new phases such as NiAl3, Ni2Al3, NiAl and Ni3Al formed and grew at the

expense of Al and Ni. The phase and microstructure development is similar to that of Ni

+ Al system. For example, at 6800C, the coexistence of the four phase

Ni/Ni3Al/NiAl/Ni2Al3 structures was also observed in the Ni + Al* system. However, the

amount of NiAl phase was smaller and more porosity was observed in the microstructures

compared with the Ni-Al samples (compare Fig. 4.16c and Fig. 4.15f). The Al203 phase

was difficult to be distinguished in SEM from porosity which also appeared black. When

the temperature was raised to 1000C, the majority phase was NiAl, with minor Ni3Al

and A1203 and some porosity. Finally, after two hours of holding at 12000C, the reaction

was complete and a nearly fully dense NiAl/a-A1203 composite was achieved (Fig. 4.16f).

4.2.4 Discussion-Reaction Mechanism

The highly exthothermic nature of the powder mixture of Ni + Al has been

utilized for reactive sintering of NiAl. Based on the Ni-Al phase diagram (Fig. 2.1), the

binary system exhibits two solid solutions and five stable intermetallic compounds,

NiAl3, Ni2Al3, NiAl, NisAI3 and Ni3Al, four of which have been detected in this study.

The reaction mechanism and phase formation sequences of the Ni-Al and Ni-Al* powder

compacts under pressureless sintering (DTA) and reactive hot compaction will be

discussed in detail below.

Pressureless Sintering. The evolution of the phases and the reaction paths during

pressureless sintering of the powder compacts may be discerned from the exothermic and

endothermic peaks in the DTA curves of Fig. 4.5 and Fig. 4.6. At high heating rate

(50C/min), only one exothermic peak was observed for the Ni-Al powder mixtures

suggesting that the formation of NiAl was achieved in one-step as supported by previous

XRD analysis [68]. However, several exothermic and endothermic peaks were observed

at lower heating rates (5-20C/min) for Ni + Al and at all heating rates for Ni + Al*. A

comparison of the temperatures for the endotherms and the melting points of the various

possible phases in the Ni-Al phase diagram (Fig.2.1) suggests that the endotherm at

-6400C corresponds to the eutectic point of Al and NiAl3, while the endotherms at

-6600C and -8600C correspond to the melting points of Al and NiAl3, respectively.

Based on the appearance of these endotherms, the first exothermic peak may be assigned


Fig. 4.15 SEM micrographs (BSE mode) of the Ni+Al powder compacts heated to (a)
4000C, (b) 4500, (c) 4500C (high magnification), (d) 5000C, (e) 6000C, (f) 6800C, (g)
7500C, (h) 8500C, (i) 8500C holding two hours, and (j) (SE mode) 12000C holding two
hours during RHC at a heating rate of 100C/min.


Fig. 4.15 Continued.

Fig. 4.15 Continued.

Fig. 4.15 Continued.

Fig. 4.15 Continued.

Fig. 4.16 SEM micrographs (BSE mode) of the Ni + Al* powder compacts heated to (a)
4500C, (b) 5500, (c) 6800C, (d) 7500C, (e) 1000C, and (f) (SE mode) 12000C holding
two hours during RHC at a heating rate of 100C/min.

Fig. 4.16 Continued.

A414 = a

0~~~~ 1 5K ie01F

Fig. 4.16 Continued.

- *

to the formation of NiAI3 while the second and the third peaks may be due to the

formation of Ni2AI3 and NiAI, respectively. Therefore, the formation sequence of the

phases during pressureless sintering of Ni + Al powder compacts may be explained as

follows. Al initially reacts with elemental nickel to form NiAI3, and this reaction occurs at

the temperature corresponding to the first exothermic peak. Upon further heating, a

portion of the NiA13 forms an eutectic with Al and melts at 6400C (first endothermic

reaction). NiAl3 reacts with Ni to form Ni2Al3 at the second exothermic peak. The 6600C

endotherm (Al melting) is not distinguishable since the second exothermic peak occurs

almost at the same temperature. The unreacted NiA13 melts upon reaching its melting

point (the 8600C endotherm). Finally, the NiAl3 and Ni2Al3 react with Ni to form NiAl

(the third exotherm).

In the case ofNi + Al* powder compacts, the 6600C endotherm (Al melting) was

clearly observed (Fig. 4.6) and the formation of Ni2A13 (the second exothermic peak for

Ni + Al) was delayed to the higher temperature (the second exotherm). In this case, the

shouldering of the second exotherm suggests the initiation of another exothermic

reaction, which may be the NiAI formation. The higher amount of heat liberated as

suggested by the area under the second peak is due to the simultaneous occurrence of

these two exothermic reactions. Thus the formation sequence of the various phases during

pressureless sintering of Ni-Al powders is possibly NiAl3 -+ Ni2A13 -+ NiAl and the

reactions corresponding to the three exothermic peaks for Ni + Al are as follows:

Ni + 3A1 -+ NiAI3 (1st exotherm, 557-587 C)

NiAI3 + Ni -* Ni2A13 (2nd exotherm, 648-653 oC)

Ni2Al3 + Ni -+ 3NiAI (3rd exotherm, 892-937 oC)

The studies of diffusion couples of Al and Ni [55-57] and the predictions of the

effective heat of formation model of Pretorius et al. [151] on the thin film compound

phase formation also indicate a similar sequence of formation. The DTA results of Dyer

and Munir [152] and Rein et al.[153] also showed three exothermic peaks at heating rates

of lower than 50K/min in the case of Ni-Al multilayers and Ni-Al powder mixtures. They

also attributed the peaks to the formation ofNiAl3, Ni2A13 and NiAl.

In contrast with the DTA results of Ni + Al powder compacts, where three

exothermic peaks are observed, Ni + Al* exhibited only two exothermic peaks with an

overlap of two exothermic reactions at the second peak. While the second exotherm

occurs at a higher temperature when oxidized Al powders are used, there is no separate

third exotherm corresponding to the formation of NiAl. This may be attributed to the

retardation effect of the oxide layer in this case. Since alumina is a strongly adhering

oxide [15], it is expected to present a barrier for reaction of aluminum with nickel in the

case when partially oxidized Al is used. Therefore, the presence of A1203 delays the

formation of Ni2Al3 in Ni + Al* powder compacts to higher temperatures which

otherwise should occur at a lower temperature of about 6500C as observed for Ni + Al.

However, at higher temperatures, the Ni2Al3 formation occurs facilitating the NiAl

formation, resulting in the overlap of the two exotherms. Atzmon's [154] investigations

on the effect of interfacial diffusion barriers like oxide layers on the ignition temperature

for self-sustained reactions also showed similar effects. It was suggested that the

interfacial barrier inhibits the reaction at lower temperature but becomes transparent at

higher temperatures. Also it was shown that for stronger barriers, the ignition temperature

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd