MSU

LIBRARIES
“

 

 

RETURNING MATERIALS:
P1ace in book drop to
remove this checkout from
your record. FINES wil1
be charged if book is
returned after the date
stamped below.

 

 

 

 

 

PHASE TRANSFORMATIONS IN (Ni,Cu)3Sn ALLOYS

BY

Jung-soon Lee Pak

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Metallurgy, Mechanics and Materials Science

1987

ABSTRACT
PHASE TRANSFORMATIONS IN (Ni,Cu)3Sn ALLOYS
BY

Jung-soon Lee Pak

The phase transformations:h1(Ni,Cu)3Sn alloys con—
taining Cu contents ranging between 10 and 22 atZ were
investigated in detail by means of optical and scanning
and transmission electron microscopy, electron and X—ray
diffraction techniques, and differential thermal analysis.
A high-temperature D03 phase was transformed to an ordered
2H (beta-CuBTi-type) orthorhombic phase at £1 temperature
of around 700°C in alloys containing Cu contents between
14 and 20 atZ. Residual areas of the D03 phase then were
transformed to both the 2H phase and a new phase at around
460°C. Electron and X-ray diffraction analyses revealed
that the new phase exhibited a triclinic structure,
slightly distorted from the 2H structure. Below that
temperature, the three phases of D03, 2H, and d2H are
present. These results were consistent with those obtained
from specimens containing 20 at% Cu, aged at various
temperatures.

Transmission electron microscope observations of
specimens with-Cu contents of 14, 17.5, and 20 at%,
furnace-cooled from 1000°C, demonstrated that the d2H
phase exhibited acicular structures containing a large

number of uniformly distributed internal faults. These

Jung-soon Lee Pak
d2H structures are surrounded by the 2H phase, which
contains rod-shape structures having (101) faults and a
low density of dislocations.

TEM observations of specimens quenched from 1000°C to
ice water revealed that the high-temperature D03 phase was
retained for alloys containing Cu ranging from 16 and 22
atZ. For an alloy with a Cu content of 14 atZ, a 2H
martensite was formed containing (121) twins.

A new phase diagram is proposed for the (Ni,Cu)3Sn
alloys in the range of Cu between 10 and 22 atZ. This
phase diagram is different from the one found in
literature, showing a single phase of 2H at temperatures
below about 900°C in the range of Cu contents between 8
and 18 atZ.]H:is also proposed that the d2H phase can be

formed from an L21 phase via the D03 phase.

ACKNOWLEDGMENTS

My sincere thanks go to Professor K. Mukherjee, my
advisor, for his stimulating presence throughout this work
and for his understanding of my being on leave from MSU to
pursue part of this work.

To Professor O.T. Inal, my coadvisor, for his
encouragement to complete this work.

To Professor G. Purcell for allowing me to use his
research facilities in the Department of Materials and
Metallurgical Engineering, New Mexico Institute of Mining
and Technology.

To Ina S.Sirkar fOr his help in the DTA experiments.

My special thanks go to my husband, Han, and my
parents, Mr. and Mrs. H. Lee for their high inspiration
and continued interest in my work.

To my son, Yune, for his stay in East Lansing with me

from the fall of 1985 to the spring of 1986.

iii

TABLE OF CONTENTS
Page
LIST OF TABLESCOOCCCOO0.00.00.00.00.00...OOOOIOOOOOCOO Vi

LIST OF FIGURESOO0.000000000000000...OOOOOIOOOOOOOOIOO Vii

CHAPTER 1 INTRODUCTION
1.1 Introduction........................ 1
1.2 Martensitic Transformations in D03
Ordered Alloys...................... 3
1.3 Massive Transformations in Cu-based
Alloys.............................. 11
1.4 The Ni3Sn System.................... 17
1.5 The (Ni,Cu)3Sn Systenh ..... n.u.n. 23
CHAPTER 2 OUTLINE OF THE PRESENT WORK.............. 26
CHAPTER 3 EXPERIMENTAL PROCEDURES.................. 28
3.1 Preparation and Heat Treatment of
Alloys.............................. 28
3.2 Optical Microscopy, X-ray Diffrac-
tion, and Scanning and Transmission
Electron Microscopy................. 30

3.3 Differential Thermal Analysisnun” 32

iv

TABLE OF CONTENTS-—continued

CHAPTER 4

CHAPTER 5

CHAPTER 6

Page

EXPERIMENTAL RESULTS AND DISCUSSIONS

4.1 Optical and Scanning Electron
Microscopy.......................... 33
4.2 Results of Diffractometer and Debye—
Scherrer Methods.................... 41
4.3 Microstructures of Furnace-cooled
(Ni,Cu)3Sn Alloys................... 58
4.4 Microstructure and Crystal Structure
of As-quenched (Ni,Cu)3Sn Alloys.n. 87
4.5 Aging Effects of the Ni-20Cu—253n
Alloy............................... 92
4.6 Differential Thermal Analysesuun"106
4.6.1 Results and Discussion.n.u.106
4.6.2 Conclusions..................110
DISCUSSION
5.1 1A Proposed Phase Diagram of the
(Ni,Cu)3Sn System...................l12
5.2 A Formation Mechanism of a Distorted

2H PhaseOOOOIOOOOOOOOCOOOO0.0.0.0...l16

SUMMARY AND CONCLUSIONS..................124

REFERENCESOOOOOOOOOOOOOOOOOOOOOOOOOOO0.0.0.00000000000127

TABLE
4.1

4.2

4.3

LIST OF TABLES

Page

Angles and intensities of reflection peaks

for a 2H-orthorhombic structure of a furnace-
cooled Ni-l4Cu-258n alloy. Lattice parameters

used for theoretical calculation: 824.510,
b-5.393, c-4.3OOA; Cu-k alpha: 1.54439OA;
Debye-Scherrer method used...................... 43

Angles, d-spacings, and intensities of

reflection peaks for a D03 structure of a
Ni-17.SCu-258n alloy. Lattice parameter used:
a-5.910A; Cu-k alpha: 1.54439A; Debye-Scherrer
method used..................................... 49

Angles and intensities of reflection peaks for

a D0 structure of a furnace—cooled Ni-l4Cu-253n
alloy. Lattice parameters used for theoretical
calculation: a=b=5.306, c-4.260A; Cu-k alpha:
1.544390A; Debye-Scherrer method usedu.u.u.u 50

Debye-Scherrer data obtained from the alloys,
furnace-cooled from 100000. Cu-k alpha used..... 55

D-SpflCiflgSOf thEdZH phaseooooooooooooooooooooo 85

Angles between two planes for d2H and 2H
StruCtureSCCOOCO00.0.00...OCOOOOOIOIOOOOIOOOOOOO 86

X-ray reflection peaks observed in a specimen
aged 24hr at 500°C..OOOOOOOOOOOOOOOOOOOO00...... 95

X-ray reflection peaks observed in a specimen
aged 24hr at 600°C....0....OOOOOOOOOOOOOIOOOOOOO 97

Summary of DTA analyses obtained from
88-quenChed (Ni,C“)38n specimen80000000000000000108

Examples of tetragonally distorted CsCl-type
compounds...‘0.0...0.0.0.0...OOOOOOOOOOOOOOOOOO.123

vi

FIGURE
1.1
1.2

1.3

1.6

1.7

1.10
3.1
4.1

LIST OF FIGURES

Page
Unit cell of D03- and LZI—type superlatticesn. 5

Two kinds of atomic layers, A and B, in (110)
planes Of D03 S121'thCUI'E..................o.....

Six kinds of atomic layers in close-packed
structures of martensite transformed from D0 -
type superlattice. The arrows indicate the
displacement vector of each layer referred to

layer A...0.0...O00....UOIOOOOOOOIOIICOIOOOOOOO 7

Unit cell of 2H (beta—Cu3Ti-type) superlattice. 8

Stacking sequence of basal planes for a 9R
structureOOCOOOOOOOOOCOOOOOOOOOOOOOIOOOOOOOOOOO 10

A portion of a phase diagram in the Cu-Ga

ayatemOOOOOOOOOOIOOOOI.00...OOOOOOOOOIOOOIOOOOO13

Lattice parameters of 5, 5', and Sm phases in
the cu-Ga system.....0......COCCCOOCODCCCOCO... 15

Phase diagram of Ni—Sn system.................. 18

Atomic arrangements in the (0001) plane of
D019 structure (a) and in the (001) plane of
2H structure (b)..........CCOOOCOCOOOCCOCCCOOCC 21

Phase diagram of Ni-xCu-ZSSn alloys............ 24
Flow chart of heat treatments employedu.n.u. 29

Microstructure of a Ni-l4Cu-258n alloy,
furnace-cooled from 1000°C to room
temperature.‘...................C.............. 34

Microstructure of a Ni-l4Cu-258n alloy,

furnace-cooled from 1000°C to room
temperature.0.0.0.0000...OOOOOOOOOOOOOOO0...... 35

vii

LIST OF FIGURES-~continued

FIGURE

4.3

4.4

4.5

4.6

4.8

4.9

4.10

4.11

4.12

Page

Scanning electron micrographs of a Ni-14Cu-258n
alloy, furnace-cooled from 1000°C.............. 37

Microstructure of a Ni-17.5Cu-253n alloy,
furnace’COOIed from 1000000....0000000000000... 39

Microstructure of a Ni-ZOCu-ZSSn alloy, furnace-
COOIedfrom10000000000000.0000...ooeoooiooooooo 40

Partial diffractometer pattern of a furnace-
COOIed Ni-l7.5Cu-ZSSD 8110,0000...0000000000000 42

Composition dependence of the lattice parameter
of D03 structure in as-quenched Ni-xCu-ZSSn

alloys...OOOOOOOOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOO 48

Partial diffractometer pattern of a furnace-
COOIed Ni-ll‘cu-zssn alloy.OOOOOOOOOOOOOOOOOOOO. 53

Partial diffractometer pattern of a furnace-
cooled Ni-20Cu-258n alloy...................... 54

Microstructure of a Ni-l4Cu-258n specimen,
furnace-cooled from 1000°C. (a) bright field
image of 2H and d2H phases. (b) diffraction
pattern, taken from the top left-hand side of

a 2H region: [111] zone axis. (c) diffraction
patterns, taken from the central region of the
micrograph: [111] zone axis for 2H and d2H.

(d) angles between two reflection spots........ 60

Enlarged micrograph of an area indicated
by the single-headed arrow in Figure 4.10...... 62

Enlarged micrograph of an area indicated

by the double-headed arrow in Figure 4.10.

Note fringes of stacking faults and

dislocation networks in a 2H region............ 63

viii

LIST OF

FIGURE

4.13

4.14

4.15

4.16

4.17

4.18

4.19

4.20

FIGURES--continued

Page

High magnification micrograph of a furnace-
cooled Ni-14Cu-253n specimen, showing many
microstructural features....................... 64

Microstructure of a Ni-l4Cu-258n specimen,
furnace-cooled from 1000°C. (a) bright field
image of 2H and d2H phases. (b) diffraction
pattern of [234] zone axis taken from 3 2H

region. (c) diffraction patterns taken from a
two—phase region............................... 68

Microstructure of a Ni-14Cu-258n specimen,
furnace-cooled from 1000°C. (a) bright field
image of 2H and d2H phases. (b) diffraction
pattern taken from an area containing bands.... 70

High magnification micrographs taken,
(a) from area A; (b) from area B; and
(c) from area C in Figure 4.15................. 71

Electron micrograph of a 2H region taken from
a Ni-14Cu-253n specimen, furnace-cooled from
1000°C. Note (101) planar faults.............. 75

Bright field image of a Ni-ZOCu-ZSSn specimen,
furnace-cooled from 1000°C to room temperature.
Note dislocations and planar faults in 2H

regions and planar faults in d2H regions....... 77

Bright field image of a Ni-200u-253n specimen,
furnace-cooled from 1000°C, showing zigzag

bandSOOOOOOOOOOOOIOOOOOOOOOOOOOOOOOOOOCOOOOOIOO 78

Bright field image of a Ni-2OCu-258n specimen,
furnace-cooled from 1000°C to room temperature.
Note many dislocation loops and dislocation

lines in 2H regions............................ 80

ix

LIST OF FIGURES-~continued

FIGURE Page

4.21 Electron diffraction patterns taken from
2H areas of Ni—14Cu-253n specimens furnace-
cooled_from 1000°C. Zone axes: (a) [100];
(b) [210]; (c) [011]; and (d) [012]............ 82

4.22 Electron diffraction patterns taken from
d2H areas of Ni-l4Cu-2SSn specimens furnace-
cooled from 1000°C. Zone axesj_(a) [001]:
(b) [010]; and (c) [001] and [214]. ........... 83

4.23 'Optical micrograph of a Ni-l4Cu-258n specimen,
quenched from 1000°C to ice water, showing
a plate-like martensite formed from a D03

phase.OCOOOOOOOOOOOOOOOOOOOOOOOOOOIOOOOOOCOOOOO 88

4.24 Electron micrograph of a 2H martensite (denoted
as 2H') formed in a Ni-l4Cu-258n specimen,
quenched from 1000°C to ice water. Note
a high density of (121) twins.................. 89

4.25 Microstructure of an as-quenched Ni-l4Cu-253n
specimen, showing two martensite variants
containing twins. (a) bright field. (b)
diffraction pattern of [120] zone of 2H, taken
from an area indicated by the single:headed
arrow. (c) diffraction pattern of [124] zone,
taken from an area indicated by the double-
headed arrow................................... 91

4.26 Partial diffractometer pattern of a Ni-20Cu-
25Sn alloy, quenched in ice water from
1000°C aged 24 hr at 650°C, and quenched in
ice water...................................... 93

4.27 Partial diffractometer pattern of a Ni—20Cu-
25Sn alloy, quenched in ice water from
1000°C aged 24 hr at 500°C, and quenched in
ice water...................................... 96

LIST OF FIGURES—~continued

FIGURE Page

4.28 Partial diffractometer pattern of a Ni-ZOCu-
255n alloy, quenched in ice water from
1000°C, aged 241urat 350°C, and quenched in
ice water...................................... 99

4.29 Diffraction pattern of a [214] zone, taken
from a d2H region in a Ni—20Cu-258n specimen,
aged at 350°C for 24 hr.0....OOOIOOOOOCCOOCOOOCIOO

4.30 Spinodal structure of a Ni-20Cu-25Sn specimen,
aged 24 hr at 600°C, following ice—water
quenching from 1000°C. (a) modulated
structures of a D03 phase. (b) two sets of
satellite reflections, perpendicular to the
modulated structures. Note no satellite of
superlattice reflections. [100] zone..........102

4.31 Phase relationship among M3Sn (MaNi,Cu,Mn)
compoundSOOOOOC0.0.0.0000...OOOOOIOOOOOOOOOOOO.104

4.32 AT-vs-T curves measured by a differential

thermal analyzer inNi-xCu—25Sn alloys

during a heating process (‘4T: temperature
difference). (a) x-14 atz: the starting phase
being a quenched 2H martensite; (b) x=17.5

atz: the starting phase being a D0 high
temperature phase; and (c) x-20 at?: the
starting phase being a D03 phase...............107

5.1 A proposed phase diagram for (Ni,Cu)33n
8110’s....O...0.0.0....0.......0000000000000000114

5.2 Atomic arrangements of closed-packed planes.
(a) the (0001) plane of D019 structure.
(b) the (001) plane of 2H structure.
(c) the (001) plane of d2H structure...........118

5.3 Schematic illustration describing a formation
process of a d2H from D03 via L21..............121

xi

1 INTRODUCTION

1.1 Introduction

Some nickel-base ordered alloys are known to show
martensitic transformations. Examples are Ni-Al[1,2], Ni-
Zn[3,4] , Ni—Zn-Cu[5], Ni3Sn[6] and (Ni,Cu)3Sn[7-9], and
Ni-Ti[10]. The Ni-Al, Ni-Zn, and Ni-Zn-Cu alloys all
undergo a martensitic transformation from a CsCl(82)-type
ordered cubic phase to an AuCuI(Llo)-type ordered
tetragonal phase. In the Ni-Al and Ni-Zn-Cu alloys,
martensitic transformations are thermoelastic, and these
alloys show shape memory effects [1,2fiiL Martensitic
transformations in Ni3Sn and (Ni,Cu)3Sn alloys occur from
a beta-Fe3A1(D03)-type ordered cubic phase to a beta—
Cu3Ti(2H)-type ordered orthorhombic phase[6]. A similar
martensitic transformation from a D03 to a 2H structure
also takes place in thermoelastic alloys such as Cu-Ni-Al
and Ni-Zn [3,4], which are known to show shape memory
effects. Interestingly, the Ni3Sn-based alloy does not
show any shape memory effect. Since the orientation
relationship and habit plane associated with martensitic
transformations are essentially the same for the Ni3Sn[6],
Cu-Ni-Al[11], and Ni-Zn[3,4] alloys, the transformation
mechanism is believed to be the same. Thus, the lack of

shape memory behavior of a martensite in Ni3Sn can not be

explained.

In the past fifteen years, extensive investigations
have been carried out to elucidate shape memory mechanisms
of shape memory alloys such as Ti-Ni, Cu-Ni-Al, Ag-Cd, and
Au-Cd alloys[12-16]. Hence, the understanding of such
shape memory mechanisms is quite good. Detailed study and
understanding of martensitic transformations in Ni3Sn
alloys is of interest, because N138n martensite is one of
the few martensites in nickel-based alloys that do not
exhibit shape memory (SM) effects. Recent studies of Ni3Sn
alloys with added Cu have demonstrated that such alloys
[have a stable 2H phase at low temperature, which has the
same structure as that of a martensite in a N135n
alloy[6]. In the literature, a new phase diagram was
proposed[7]. After detailed investigation of the reported
data, however, it was found that some phenomena could not
be understood by the explanation given in the literature.
Therefore, in the present case, we decided to study phase
transformations in (Ni,Cu)3Sn alloys to obtain a correct
phase diagram for the alloy system“ A further reason for
choice of this alloy system is that the phase
transformations of ordered nickel-based alloys have not
been studied as extensively as in Cu-base ordered alloys
where martensitic transformations commonly manifest shape

memory effect.

1.2 Martensitic Transformation in D03 Ordered Alloys

Mostbeta phases in noble-metal alloys, with a 3:2
electron-to-atom ratio, e/a, have an ordered bcc structure
at high temperatures. The electron-to-atom ratio is
defined as the average number of free electrons per atom.
According to the electron theory of alloys, a bcc
structure is considered to be stable at a value of
e/a~1.5, when a monovalent fcc noble metal ( such as Au,Cu
etc. ) is alloyed with a bivalent metal ( such as Zn,Cd
etca L.As the number of free electrons increase with the
addition of the bivalent metal, the density of available
states decreases in the fcc structure than in a bcc
structure. The peak in the density of state in the bcc
phase occurs at about e/a-l.48. Hence, the energy of the
conduction electrons is lowered in the bcc structure at an
electron-to-atom ratio of 1.5 compared with a fcc
structure at the same e/a ratio[17].

Such a beta phase mostly has a fairly wide composition
range of solubility at high temperatures. The stability
of the beta phase decreases with decreasing temperature,
resulting in a narrow range of solubility at lower
temperatures. Thus, the beta phase usually decomposes
below eutectoid temperatures of about 700°C. If,<u1the
other hand, the beta phase is rapidly quenched to suppress

the diffusion of atoms, the metastable beta—phase

transforms martensitically at a low temperature. Crystal
structures of such transformation products are mostly
close—packed layer structures. Because parent phase has
an ordered structure, the martensite product phase
inherits ordering since the transformation is
diffusionless. Typical crystal structure of beta phases in
noble-metal alloys manifest eitheran Fe3Al(D03)-type
superlattice(Figure1.1)ora CsC1(B2)-type superlattice.
Both are called ordered bcc structures, although it should
be noted that strictly speaking, the D03 and B2 structures
can be interpreted as fcc and simple cubic structures,
respectively.

The D03 structure is considered to be composed of two
kinds of atom planes, A and B, that are alternately
stacked parallel to the (110) close packed plane of the
bcc structure (Figure 1.2). Martensitic structures are
considered to be formed by a shear motion of atoms on the
(110) planes of the bcc structure. It is possible to form
six different types of close-packed layers by shear
motions parallel to [1T0] on (110), although they are
shifted relative to each other in the [lIO] directions.
Such possible planes are shown in Figure 1.3, where the
prime of A', B' and C'represents a change in atom arrange-
ment of the A, B, and C planes, respectively, by b/2 along
the b-axis. For instance, the gamma-prime 2H structure

(Figure 1.4) of N138n has an AB' stacking order. For a 9R

 

 

 

 

 

 

        

 

 

 

Q

 

 

 

 

 

 

 

 

 

 

 

 

 

a

Quunu
1:- x

:

:

:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. b-site

O a-site
@ c-site Q d—site

Ni3Sn(DO3) Ni:@®., SnzO
CuzNiSn (L21) Cu:@®, Ni:., Snzo

Figure 1.1 Unit cell of D03- and L21-type superlattices.

.(110)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

(110)
. Fe 0 Al

Figure 1.2 Two kinds of atomic layers, A and B,
in (110) planes of D03 structure.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

f7"
A
MW;

 

 

 

A’ B'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E

v
“l||[ .
W

.Ni 0 Sn

Figure 1.3 Six kinds of atomic layers in
close-packed structures of
martensite transformed from D03-
type superlattice. The arrows
indicate the displacement vector
of each layer referred to layer A.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2H ( B-CU3Ti-type )

Figure 1.4 Unit cell of 2H (beta—Cu3Ti-type)
'superlattice.

structure (Figure 1.5), three layers constitute one
subperiod. But, if atomic ordering is taken into
consideration, the six layers AB'ACUHT constitute one
subperiod. Thus, the structure is called a 18R. If this
subperiod stacking is taken as the unit cell, the
resulting crystal type is a monoclinic structure. If
nine layers AB'CB'CA'CA'B (so-called 9R) are taken as the
unit cell, the crystal type is then an orthorhombic
structure. The a- and b- axes in the orthorhombic
coordinate system are shown in Figure 1.5: the c- axis is
perpendicular to the close-packed plane.-

Sato et al [18,19] explained the reason for the
existence of the layer structures in terms of the electron
theory of alloys. If stacking faults are introduced
periodically into a crystal, the crystal has a long—period
stacking order, resulting in the fOrmation of a: new
Brillouin zone boundary. If the electron-to-atom ratio
happens to be such that the Fermi surfaCe is almost in
contact with the newly formed Brillouin zone boundary, the
gain energy of conduction electrons becomes greater than
the increase in strain energy accompanied by the
introduction of stacking faults at regular intervals.
Thus, such long-period stacking structures become stable.
This, however, needs atomic shuffling in addition to the
shear motion of atoms.

Because energy differences among various kinds of

10

 

 

 

Figure 1.5 Stacking sequence of basal planes
for a 9R structure.

11

long-period stacking structures are believed to be small,
there are other factors, in addition to the alloy
composition, that determine which long period stacking
structure will be formed. In Cu-Al alloys, 8 martensite
forms in the bulk specimen which has the 9R structure,
while in a thin foil of the same alloy, another
martensite having the 2H structure is formed[20]. In some
other alloys, a mixture of two kinds of long-period
stacking structure is formed. ,An example is the Au~Cd
system, where the 2H and 9R structures are found in a
lamellar form[19]. Factors to determine which long-range
stacking structure should be present in an alloy have not

yet been clarified.

1.3 Massive Transformations in Noble-Metal Alloys

In 1939, Greninger[21] found a patchy microstucture
in a Cu-9.3 atZ Al alloy with a fcc crystal structure. He
designated this microstructure\as "massive structure".
This massive structure was obtained by quenching the alloy
from a temperature range where the bcc high-temperature
phase is stable. The massive structure showed equiaxial
grains, and the grain boundaries were mostly straight with
some irregularity but not oriented in specific directions.
Essential differences between a precipitate and a massive

structure are as follows; 1) the latter is not formed by

12

long-range diffusion of atoms and consists of the same
alloy composition as the parent phase; that is, the
massive transformation is a diffusionless one; 2) the
precipitation takesplacebylong-rangediffusionof atoms,
resulting in compositional changes. More detailed descrip-
tions of the massive transformation will be given below.
Since 1934, massive structures have been found in many
other Cu alloys. These alloys are Cu—Ga, Cu-Zn, Cu-Zn-Ga
and Cu-Ga-Ge alloys, all of which were studied by Masalski
et al[22-24]. Saburi and Wayman[25] also found a massive
structure in a Cu-23.8 atZ Ga and studied the
microstructure in detail by means of optical and
transmission electron microscopy. They obtained the
massive structure by quenching the beta phase of the alloy
into warm water to avoid a eutectoid reaction at 610°C.
It was found that this massive structure had the following
morphological features: irregular or straight massive
grain boundaries, a feathery structure, and a lamellar
structure. The morphologies of the massive structure,
thus, appear complex in form. The massive phase has a
crystal structure and unique morphologies, different from
those of a 18R martensite obtained by quenching specimens
of the same alloy into ice water. An equilibrium phase
diagram of the Cu-Ga system [23,26]is shown in Figure 1.6.
Masalski[22] also studied a massive structure of Cu-Ga

alloys in detail and determined the crystal structure of

l3

 

900 ‘

\ \\
800 \
l3
x Z
a \__2_2.4 616'
500

23.7 r
F. _q__ _ ’l
\ M. (3” (a) I,
\ I
500 bfi._--v--qp--
‘\
/ [s 71

\
40° II M. (la—~30
--4.--»-- --- ---1‘

300 U '4’ ‘fi ‘
l 1

18 20 22 24 26 28 30 32
Ga (n96)

Figure 1.6 A portion of a phase diagram in
the Cu-Ga system.

 

 

 

 

 

( °C)

 

 

 

 

 

 

 

 

 

 

 

 

Temperature

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

14

this phase (zeta-m phase) by X-ray diffraction. (The m
herein indicates the phase formed by a massive
transformation.) The structure determined (M? the zeta-m
phase is a hcp structure, the same as the equilibrium
zeta and zeta—prime phases. The lattice parameter of
zeta-mphaseincreaséslinearly with Ga(Figurel.7) [22].
This feature suggests that the zeta-m phase is not a
decomposition product but a supersaturated solid solution
of the zeta or zeta-prime phase. As can be seen from the
phase diagram (Figure 1.7), a change in the lattice
parameter, a, of composition of the zeta-m phase is the
-same as that for the beta parent phase. It is, hence,
iconcluded that the massive transformation is governed by
a diffusionless mechanism.

Significantly, the diffusionless character of this
transformation is the same as that for martensitic
transformations. However, there are some different
features between massive and martensitic transformations.
Kittl [26] studied the effect of quenching rates on Ma
temperature (I temperature at which a, massive
transformation starts to take place}. He found that the
Ma temperature of Cu-Ga alloys did not show any
considerable change when the quenching rate was changed
between 10 and 2500°C/sec. It should be noted that the Ma
temperature was higher than the Ms temperature (at which

martensitic transformations start to take place) of the

15

 

 

 

 

4'“,
I
2.605 ~4250 ,’ 1.635
r' ,I
\i ,
a C [I
o o I a
E
g \\x
H
a 2.600 ”-4.240 I 1.630
a, \
”:1 \V \
5 \‘\c \‘Q~
J \
\
c/a‘
\
'§
2.595 —4230 1.625
1

 

 

 

 

 

 

21 22 23' 24
Ga (21%)

Figure 1.7 Lattice parameters of C, C', and Cm phases
in the Cu-Ga system.

16

alloys.

Kittl [26] also found another peculiar characteristic
in the zeta-m phase. That is, there is no unique
orientation relationship between the beta matrix phase
and the zeta-m phase. Grains of the zeta-m phase are
often seen to have grown across grain boundaries of the
parent phase[23]. Such phenomena are not observed in the
case of martensitic transformations. In this case,
specific orientation relationships are present and no
martensites can grow across grain boundaries of their
parent phase.

Transmission electron microscopy and electron
diffraction techniques showed that there were numerous
internal faults in the zeta-m phase. These internal
faults are parallel to the (0001) basal plane of the zeta-
m(hcp) phase. It is interesting that the number of faults
is less in the zeta-m phase than in the martensite. When
quenching rate is very low, only a.few faults are formed
in the zeta-m phase. This lesser number of internal
faults in the massive structure suggests that the massive
transformation occurs with some sort of diffusion of
atoms. From observations of microstructural features and
changes in lattice parameters, however, this diffusion
appears to be short-range diffusion, rather than long-

range diffusion.

17

1.4 The Ni3Sn System

Anequilibrium phase diagram of the binary Ni-Sn
system was shown in Figure 1.8 [27]. The Ni3Sn alloy has
a D019 ordered hexagonal structure at low temperatures and
a D03 ordered structure at high temperatures. Because a
phase transformation to the D019 phase was not clear,
Niel [28] studied the crystal structure of the D03 high
temperature phase. He reported that X—ray data obtained
from a quenched powder specimen showed a disordered hcp
structure, and concluded that a transformation to the D019
phase was an order-disorder one. This conclusion was not
consistent with X—ray data of the high temperature phase
reported by Schubert et a1 [29]. They used high
temperature X-ray equipment and determined the crystal
structure to be a D03 structure. The D03 structure at
high temperature was strongly supported by a study carried
out by Pak et al [6].

Pak et al [6] studied phase transformations of the
high temperature phase by detailed microstructural
observations and identification of phases present in
specimens. The specimens used were quenched from 1050°C
(a temperature region of the D03 phase) at several
different cooling rates. Methods of investigation
included metallography, electron microscopy, and electron

and X-ray diffraction techniques. The reported results are

18

IIIGHT PER CENT YIN

 

10 2° 30 50 55 60 65 70 75 60 65 90 95
'500 l I i l J J __l I I I I
“53'

 

 

 

*—-Ni, 30.

«001‘
\ \
\
\
1300 :— A.
\ 1264'
\‘ I
I
1200 I

 

 

 

 

 

 

\
‘7’ O
\ "I. "50‘;
V "30°
:13 ]
uoo ——,',33,‘ (32.5:

 

 

 

mu 4 l \
921' . \

050°
7”. i559.” '——‘

\
:202' \

600 .4 -—_... ..

/ 1
500 '
co \

 

§

TEMPERATURE, 'C
§
-\
\\

 

 

 

N
O
O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30¢ \umw. .
' must. 2“.
’
. 99.: '32.
200 V (”.851
|
800 \
0 IO 20 30 4O 50 60 70 80 90 100
ll ATOMIC PER CENT III SI

Figure 1.8 Phase diagram of Ni-Sn system.

19

described as follows.

A massive transformation was found to have taken place
by quenching from 10500 to 0°C at cooling rates of
10°C/hr, 40°C/sec and 200°C/sec. Grains of the massive
phase crossed original grain boundaries of the parent
phase during their growth, similar to observations made
of other massive phases. The crystal structure of the
massive phase in the N138n was a D019 ordered hexagonal
structure, the same as that for the stable phase at low
temperature. Microstructures of the massive phase are
different from those of the stable low-temperature phase,
although both the phases have the same crystal structure.
The massive phase contained planar faults parallel both on
the basal (0001) plane and on the prism (IOIO) plane,
while the stable phase did not contain such planar faults
but contained a few grown-in dislocations.

0n the other hand, when the high temperature D03
phase was quenched at a rate of 200°/sec or higher, very
fine acicular structures were formed. 'These structures
clearly showed surface relief, and thus it was concluded
that they were martensites. The crystal structure of the
acicular martensite was determined to be a 2H ordered
orthorhombic structure (beta-CuBTi-type), which may be
calledadeformed hcp structure. Pak et al [6] namedthis
martensite phase the gamma-prime phase. This gamma-prime

martensite contains (121)2H internal twins, which provide

20

the lattice invariant strain. By taking into account
atomic arrangements oftflua(0001) massive plane and the
(001) martensite plane, and by also taking into account
[lOIO] planar faults in the massive phase, a relationship
between the D019 massive phase and the 2H martensite phase
can be determined. As seen in Figure 1.9, which shows
atomic arrangements for these two phases, it is possible
to produce planar faults by applying an a/2 <1210> shear
on every other[1OIO}plane [6]. Hence, planar faults
observed in the massive structure are most probably thin
crystals of the 2H martensite phase. Until recently,
because oftidifficulty of retaining the parent phase of
D03 structure at room temperature, no orientation
relationship between the 2H martensite and the D03 parent
phase was reported.

Recently, Chang [30] proposed a new method of
determining the orientation of the parent phase and the
habit plane normals of :1 plate-shaped product phase,
without the need for retaining the parent phase at
observation temperatures. This method only requires the
sufficient number of traces formed by the intersection of
product phase variants with a surface of a specimen.
Such a surface is assumed to be the surface of the parent
phase. Pak et al[31], more recently, developed the model
proposed by Chang[30] and applied their model to the

martensite in the Ni3Sn alloy.

21

 

Figure 1.9 Atomic arrangements in the (0001) plane of
D01 structure (a) and in the (001) plane
of 2H structure (b).

22

The habit plane of a martensite in Ni3Sn was
determined to be [133}. It.is to belnentioned that a 2H-
type martensite in a Cu-Al-Ni alloy, formed from a D03-
type structure of the parent phase, has a shape-memory
effect and the martensite, has the same habit plane as the
martensite in the Ni3Sn alloy. In addition, the
martensite in the Cu-Al-Ni and Ni3Sn alloys both have
[121] internal twins. It is interesting, however, that
the latter alloy does not show any shape-memory effect.
This suggests that a thermodynamic approach is needed in
the case of martensite in Ni3Sn to understand why this

martensite does not show any shape memory capability.

23

1.5 The (Ni,Cu)3Sn System

As seen in the phase diagram for the Ni-Sn system
(Figure 1.8) the high temperature D03 phase is transformed
at around 900°C. Since this transition temperature is high
Murakamiet a1 [7—9] attempted to make it lower by
substituting copper atoms for nickel atoms. They studied
phase changes in several Ni—xCu-ZSSn alloys (x-atomic
percent) by means of X-ray diffraction, differential
thermal analysis, and optical and electron microscopy. A
new phase diagram of the ternary Ni-xCu-ZSSn system
(Figure 1.10), was proposed based on data obtained by
Murakami et a1 [7-9]. It was reported that a phase
transformation from a D03 structure to s 2H structure and
that from the 2H structure to a D019 structure were
observed by means of electron microscopy in the Ni-4Cu-
25$n alloy.

As seen in Figure 1.10, three phases with D03, 2H and
D019 structures were stable at room temperature, while in
the binary Ni-Sn system the D019 hexagonal phase is the
only phase at room temperature. In a range of copper
contents up to 6.5 at.Z,two transformations were found.
One is a martensitic transformation from the D03 phase to
the 2H phase, taking place at high temperature upon rapid
quenching, and the other is a massive transformation from

the D03 phase to the D019 phase, taking place upon slow

24

 

 

 

 

 

 

 

 

T‘I‘cr 003
IIOO:?>,‘3I‘~‘\ .
\‘QQLS ‘~.,
900b d
8
,_ 700-
500 .
r 00 H
2H
zoo . L 4 . 4
o 4 a 12 '6 20 24
Cu 01. 70
Figure 1.10 Phase diagram of Ni-xCu-ZSSn alloys.

25

cooling. In a range of copper contents from 7.5 to 18
atZ, the D03 phase undergoes a martensitic transformation
to a phase with the 2H structure upon rapid quenching.

Murakami et a1 [8] also reported that for this range,
a low-temperature 2H phase was massively transformed upon
furnace cooling. Microstructural features of the massive
2H phase are different from those of the martensite 2H
phase. Stacking faults and other planar defects are
considerably less abundant in the massive phase than in
the martensite phase. By taking structural observations
[into account, it was speculated that if the cooling rate
becomes slower, the morphology could gradually change from
a martensite-type structure to a massive-type structure by
the migration of interfaces upon cooling.

The following point is important and thus should be
emphasized. According to the phase diagram proposed by
Murakami et al[7-9], the D03 parent phase and the massive
and/or martensite 2H phase should be retained at room
temperature in a range of copper contents between 18 and
21 atZ. Thus, one can expect that some crystallOgraphic
information, such as orientation relationships between the
martensite and the parent D03 phase, or those between the
massive and the parent phase could be obtained. No such

relationships have yet been observed directly both for

Ni3Sn and for the (Ni,Cu)3Sn alloys.

2 OUTLINE OF THE PRESENT WORK

As pointed out in the previous section, if the phase
diagram of the Ni-xCu-ZSSn proposed by Murakami et a1 is
correct, then the higher temperature D03 phase could be
retained in the range of copper contents, x, between 18
and 21 atZ. Hence, the orientation relationship between
the D03 parent and the product phases should be readily
clarified for both the binary and ternary Ni-xCu-ZSSn
alloys. Attempts to produce two-phase specimens were
made. Surprisingly, however, no D03 phase could be
obtained by changing both cooling rates and Cu content.
This indicates that the phase diagram proposed by Murakami
et a1 [7-9] is incorrect. The present work, thus,
primarily focuses on clarifying phase transformations
which occur in ternary Ni-xCu-ZSSn alloys. As a further
result of the present study, a new equilibrium phase
diagram is proposed. Various aspects of this investigation
are itemized as follows:

(1) X-ray diffraction experiments were conducted to
reexamine the phase diagram of the Ni-xCu-ZSSn alloys (x
ranging from 5 to 22). Since some uncertainties remained
in the range of copper contents between 10 and 202,
special attention was paid to phase changes occurring in
this composition range. Crystal structures were determined

by a Debye-Scherrer method.

26

27

(2) As mentioned in the previous section, the 2H
structure in the (Ni,Cu)3Sn alloys was obtained by
furnace cooling, and was reported to be a stable,
equilibrium phase. It was also reported that this stable
phase was only present in the range of copper compositions
from 7.5!3)18 atl<hh In the present study, however, it
was found that there were several X-ray diffraction peaks
that were obtained in furnace-cooled specimens. These
peaks could not be indexed by any familiar structures in
this alloy system, such as 2H, D019, and D03 structures.
These results suggested the presence of a previously
unrecognized new phase. Therefore, careful, detailed
microscope observations were made in order to confirm the
presence of a new phase in this region.

(3) Transmission electron microscopy and electron
diffraction analysis were employed extensively for
identification and characterization of the new phase which
was found.

(4) Based upon results obtained through electron
diffraction analyses and electron microscopy, a new phase
diagram is proposed. Results obtained by differential
thermal analysis were incorporated into this phase
diagram.

(5) Detailed microstructural observation of phases

formed under various conditions were made.

3 EXPERIMENTAL PROCEDURES
3.1 Preparation and Heat Treatment of Alloys

Several ternary alloys were prepared from pure nickel
(99.9992 purity), electrolytic copper (99.99%) and pure
tin (99.9992). These pure metals were encapsulated in
quartz tubes (18mm diameter) after evacuation to about
10-3MPa. Encapsulated pure metals were then alloyed in a
resistance furnace and stirred vigorously several times to
produce alloys as homogeneous as possible. Weight loss was
less than 0.5% after alloying. Out of several alloys
produced, the three alloys, Ni-l4Cu-25Sn, Ni-l7.5Cu-25$n,
and Ni—20Cu-258n, were extensively investigated.

The alloy ingots were again encapsulated in evacuated
quartz tubes and homogenized for two days at 1000°C, and
subsequently air-cooled. Because all the alloys produced
were brittle, none could be rolled. The ingots were,
therefore, sliced with a diamond blade and slices (about
0.5mm thickness) were then encapsulated in evacuated
quartz tubes to anneal them at 1000°C for 2 hr. Then
various heat treatments were performed. The heat
treatments employed are listed in Figure 3.1,:and their
purposes are given as follows:

[1] After annealing for 21urat 1000°C, some specimens

were furnace cooled to investigate crystal structures and

28

29

Heat-treatments employed in this study

 

 

[Homogenization at 1000°C]

Furnace cooling to‘ Observation]
J, room temperature 7 r '
, i -
[Ice water quenching] >[Observationj

 

 

 

 

 

 

 

IAging at 350, 500, 600, and 650°C for 24hr]

 

 

 

 

[Ice. water quenching} ciflObservation

 

 

Figure 3.1 Flow chart of heat treatments employed.

30

microstructures of phases present.

[2] Annealing for 1 hr at 1000°C , followed by quenching
into ice water (the quartz capsules were crushed during
quenching), was made for the observation of the D03 and
martensite phases.

[3] Some quenched specimens (of Ni-ZOCu-ZSSn alloy) were
aged at 350, 500, 600, and 650°C for 24 hr to determine

both phase boundaries and transformation temperatures.

3.2 Optical Microscopy, X-ray Diffraction Analysis, and

Scanning and Transmission Electron Microscopy

Thin foils for transmission electron microscopy were
made with a twin-jet Fishione electro-polisher and a
conventional polisher. The polishing solution was a
mixture of 75volZ ethanol and 25volZ perchloric acid, and
polishing was done at -60°C. Transmission election
microscope observations were made with Hitachi H-800 and
HU-200F electron microscopes, both of which were operated
at 200keV.

Powder samples were prepared for X-ray Debye-
Scherrer and diffractometer methods. The powder samples
were encapsulated in a way similar to that mentioned above
and then heat-treated. In this case, the capsules were not
crushed. Cu radiation was used for all X-ray experiments.

Specimens for optical microscope observation were

31

electro-polished following mechanical polishing. The
electro-polishing solution used was the same as that for
transmission electron microscopy. After electro-polishing,
the specimens were kept for about 1 min in the same
solution for slight etching. For optical microscopy, a
Nikon microscope with a Nomarski-type interference
capability was employed. For observation of surface relief
for fine structures formed during furnace cooling, 3
Hitachi HHS-2R scanning electron microscope was used.

This electron microscope was operated at 30 keV.

32

3.3 Differential Thermal Analysis

In order to obtain the temperature of phase transi-
tion, a DuPont Thermal Analyzer (model 990) was used.
This analyzer was designed to provide two kinds of data
which are obtainable by differential scanning calorimeter
(DSC) and differential thermal analysis (DTA) methods. In
the present study, we used data obtained by the latter
method because higher temperatures could be achieved in
this mode. In the DTA method, the transition temperature
is given by an intersection of a pair of the tangents to
concave or convex DTA response curves corresponding to an
exothermic or an endothermic reaction, respectively.

In each experiment, a specimen of approximately 70 mg
was used and a heating rate of 10°C/min was employed in
all experiments. During the experiments, dry argon gas was
passed through the furnace tube of the DTA to maintain an
inert atmosphere. Alloys prepared in this study were ones
containing Cu contents of 14, 17.5, and 20 atZ, and these
specimens were quenched from 1000°C to ice water prior to

DTA experiments.

4 EXPERIMENTAL RESULTS AND DISCUSSIONS

4.1 Optical and Scanning Electron Microscopy

Surface observations of microstructures in the Ni-
l4Cu-258n alloy were carried out by means of optical
microscopy and scanning electron microscopy. Figure 4.1
shows a typical micrograph of a Ni-14Cu-258n specimen,
which was annealed at 1000°C to produce the high
temperature beta (D03) phase and then slowly cooled in an
electric resistance furnace. From the morphology seen in
Figure 4.1, it is obvious that at least two phases were
co-existing under the above-mentioned conditions. The
cooling rate in this case was estimated to be
approximately 20°C/min in the temperature range from 10000
to400°C.Thus,allphases obtained areconsidered stable.

As seen in the photomicrograph in Figure 4.1, the
most abundant phase has a taper-ended, rod-shape
microstructure. This phase appears as bright bands in the
figure. The width of these bands varies from 10 to 30 pm.
Figure 4.2 shows a similar structure obtained from
another area of the same specimen which reveals more
details about this rod-shaped structure. It is seen to
consist of fine microstructures, which bear resemblance to
the midribs of a ferrous martensite. The rod-shaped phase

was identified as a 2H ordered structure by X—ray and

33

34

Figure 4.1 Microstructure of a Ni-14Cu-258n alloy,
furnace-cooled from 1000°C to room
temperature.

    

  

J‘ I ‘
'. v
v , ‘7. '
i '
. I h
. v

r715 Q"

' ' 3%. 1:...4 100wm
I l . . . , —

   
 

35

Figure 4.2 Microstructure of a Ni-14Cu-253n alloy,
furnace-cooled from 1000°C to room
temperature.

 

36

transmission electron diffraction analyses, as will be
described below.

The less abundant phase is seen between bands of the
afore-mentioned rod-shaped phase. It appears as dark
areas in figures 4.1 and 4.2. Upon close examination by
scanning electron microscopy, it was found that the dark
areas are comprised of numerous, very fine acicular
structures. Figures 4.3(a) and (b) are scanning electron
micrographs of a specimen which was electro-polished
immediately prior to observation. From these photos, it
is seen that the acicular structures (which are surrounded
by the 2H, rod-shaped structures) appear to have specific
orientations to the parent phase. Taking this alignment
of such acicular structures into account, one can
conclude that they nucleated in the parent phase and grew
in such a way as to have a specific orientation
relationship with the original matrix of the D03 phase.

Determination of this orientation relationship would
be of interest. However, in this study this was difficult
because no parent D03 phase could be retained by furnace
cooling. An attempt was made to apply a method developed
by Pak et al [31] which was used for identification of the
habit plane of plate-like microstructures. Unfortunately,
the method was found not to be applicable in this case,
because there were multi-phases in one grain.

Similar acicular structures were observed in two

37

Figure 4.3 Scanning electron micrographs of a
Ni-14Cu-258n alloy, furnace-cooled from
1000°C.

 

38

other alloys, Ni-17.SCu-253n and Ni-20Cu-253n. Figure 4.4
shows a typical example of acicular structures that were
obtained in a Ni-l7.SCu-25Sn specimen furnace-cooled from
1000°C to room temperature. Another example is shown in
Figure 4.5, which was obtained from a furnace-cooled Ni-
20Cu-258n specimen. Ihi these figures, In: rod-shaped
structures were noted by optical microscopy. However, it
is obvious that two phases are present in both cases.
One phase appears to have a wider and longer plate shape,
while the other has a narrower and shorter plate shape.
To obtain further details about these phases, X-ray
diffraction experiments were carried out for the
previously mentioned three alloys; the results are

discussed in the following section.

39

Figure 4.4 Microstructure of a Ni-17.5Cu-258n alloy,
furnace-cooled from 1000°C.

 

40

Figure 4.5 Microstructure of a Ni-20Cu—253n alloy,
furnace-cooled from 1000°C.

 

4.2 Results of Diffractometer and Debye—Scherrer Methods

Figure 4.6 shows the results of an X—ray diffracto-
meter examination of a powder specimen of the Ni—17.5Cu-
253a alloy, furnace-cooled from 1000 (NJ to room
temperature. Becausean orthorhombic 2H stacking ordered
structure was reported to exist both in quenched specimen
of the Ni3Sn alloy [6] and the Ni-xCu-ZSSn (x-4-8) alloys
[7-9], we first tried to index reflection peaks obtained
by assigning the 2H structure. As shown in Figure 4.6,
most of the peaks could be indexed by reflection indices
of the 2H structure. The angles and d-spacings measured
allowed determination of lattice parameters of the 2H
structure as a-4.495, b-5.380, and c=4.285A. These values
are in agreement with those reported by Pak et a1 [6] and
Murakami et al [7,9].

The theoretical intensities of reflections were also
calculated and listed in Table 4.1. From this table, it
can be found that there are no strong peaks between
(121)2H and (002)2H. Thus, peaks 5 and 6 in Figure 4.6
could not be indexed by any reflection indices of the 2H
structure, indicating the presence of another phase. This
is inconsistent with the phase diagram proposed by
Murakami et a1 [7]. One of these unidentified peaks

probably can be indexed by a reflection index of a D03

41

42

57.5N1417.50u-2ssn

 

3
Angle (‘2 9 )

Figure 4.6 Partial diffractometer pattern of a furnace-
cooled Ni-l7.5Cu-25$n alloy.

43

Table 4.1 Angles and intensities of reflection peaks
for a 2H-orthorhombic structure of a furnace-
cooled Ni-14Cu-258n alloy. Lattice parameters
used for theoretical calculation: a=4.510,
b=5.393, c=4.300A; Cu-k alpha: 1.544390A;
Debye-Scherrer method used.

H K L D-SPAC INTEN L-P MULTI SF THETA
l 0 0 4.5100 1. 65.304 2 1 9.86
l 1 0 3.4597 8. 37.286 4 111 12.90
0 l 1 3.3621 11. 35.062 4 1111 13.28
1 0 1 3.1122 7. 29.660 4 111 14.37
0 2 0 2.6965 18. 21.622 8 1111 16.64
1 1 1 2.6955 5. 21.604 8 1 16.65
1 2 0 2.3144 56. 15.287 4 9 19.49
2 0 0 2.2550 27. 14.395 2 9 20.03
0 O 2 2.1500 103. 12.880 2 9999 21.05
2 1 0 2.0805 5. 11.920 4 111 21.79
1 2 1 2.0379 290. 11.351 8 999 22.27
2 O 1 1.9970 141. 10.816 4 999 22.75
1 O 2 1.9408 2. 10.100 4 1 23.45
2 1 1 1.8728 3. 9.268 8 l 24.35
1 1 2 1.8261 9. 8.718 8 111 25.02
2 2 0 1.7298 1. 7.637 4 l 26.51
0 2 2 1.6811 5. 7.118 4 1111 27.35
1 3 0 1.6699 4. 7.002 4 111 27.54
0 3 1 1.6586 5. 6.885 4 1111 27.75
2 2 1 1.6048 8. 6.348 8 111 28.76
1 2 2 1.5752 66. 6.062 8 9 29.36
1 3 1 1.5566 2. 5.887 8 1 29.74
2 O 2 1.5561 32. 5.882 4 9 29.75
3 0 0 1.5033 2. 5.403 2 1111 30.91
2 1 2 1.4951 7. 5.330 8 111 31.10
2 3 0 1.4057 3. 4.589 4 111 33.32
0 1 3 1.3852 4. 4.432 4 1111 33.88
3 1 1 1.3724 8. 4.335 8 1111 34.24
1 0 3 1.3660 3. 4.288 4 111 34.42
0 4 0 1.3483 208. 4.159 8 9999 34.94
2 2 2 1.3478 2. 4.155 8 1 34.96
2 3 1 1.3361 2. 4.073 8 1 35.31
1 1 3 1.3242 2. 3.990 8 1 35.67
1 3 2 1.3188 6. 3.954 8 111 35.84
3 2 0 1.3131 100. 3.915 4 9999 36.02
1 4 0 1.2918 1. 3.775 4 1 36.71
1 4 1 1.2371 5. 3.446 8 111 38.62
3 O 2 1.2320 3. 3.417 4 1111 38.81
1 2 3 1.2186 137. 3.344 8 999 39.32
2 0 3 1.2097 68. 3.297 4 999 39.67
2 1 3 1.1803 2. 3.153 8 1 40.86

#NUWONU‘HWth§HNNbHNNHb§ONw#UNHH§HOH§OH##NHUDNUDO#ONNI

CMHW¢§CU#NUNU\U\HQ#OU‘NHONQ#NHbun-“NOWmHowOHwa3~NWOF#U-1N

UJNOUDOUO#N&HNNH§OU#Crl-‘NNFWOHuNHbO#HOkat-‘OWNHI—NWONONI“

D-SPAC
1.1765
1.1572
1.1422
1.1275
1.1207
1.1206
1.1174
1.1139
1.1073
1.1037
1.1036
1.0906
1.0876
1.0750
1.0690
1.0490
1.0462
1.0457
1.0402
1.0266
1.0191
1.0190
1.0187
1.0111
1.0037
1.0036
0.9986
0.9985
0.9818
0.9750
0.9730
0.9704
0.9596
0.9552
0.9550
0.9490
0.9428
0.9364
0.9324
0.9131
0.9095
0.9039
0.9020
0.9004
0.8988
0.8985
0.8896
0.8865
0.8862

INTEN
22.
88.
ll.

3.

174.
130.

§ 0‘
NuNLfiNU‘NN
o o o o o o o o

b

&
HQONNN¢N©QG§W®GM§N#UIVUINUli-‘P-‘UJUJ

N

N

44

L-P
3.136
3.051
2.991
2.937
2.914
2.913
2.903
2.891
2.871
2.860
2.860
2.825
2.817
2.789
2.777
2.746

2.742.

2.742
2.737
2.729
2.728
2.728
2.728
2.730
2.734
2.734
2.739
2.739
2.765
2.781
2.786
2.793
2.828
2.845
2.845
2.871
2.902
2.937
2.961
3.106
3.138
3.193
3.213
3.230
3.247
3.251
3.359
3.402
3.406

MULTI

mookoooooomoooooooooooooooomooaboooomaooboooooooooobbbaoowoobsxooooooooooSN:>:.~oo

1111
111

999
1111
1111

111

111

999

THETA
41.02
41.86
42.53
43.23
43.55
43.56
43.71
43.89
44.22
44.40
44.40
45.07
45.23
45.92
46.25
47.40
47.57
47.60
47.93
48.78
49.26
49.27
49.29
49.80
50.29
50.30
50.65
50.65
51.86
52.38
52.52
52.73
53.58
53.94
53.95
54.46
54.99
55.55
55.91
57.75
58.11
58.68
58.88
59.05
59.22
59.25
60.23
60.59
60.62

OJ-‘NkNUINLnbbNNHmr—NmHOLnuHNLnOJ-‘HHJ-‘ONMwOHLSUiJ-‘wkwu:

U00MJ>WHNW§OLDNWO¢HJ>OONHO‘N#NOUI#HuNUIUO‘bl-‘UOHO‘ON

U1>NO©HMNWNUIDJUIONHN§NN§UOHJ>WUOUHUFOHWHOHN#UOF‘E"

D-SPAC
0.8828
0.8815
0.8745
0.8744
0.8729
0.8712
0.8649
0.8635
0.8618
0.8587
0.8554
0.8539
0.8493
0.8479
0.8465
0.8448
0.8419
0.8405
0.8390
0.8349
0.8346
0.8318
0.8318
0.8293
0.8263
0.8220
0.8196
0.8156
0.8062
0.8061
0.8050
0.8035
0.8024
0.7948
0.7948
0.7948
0.7924
0.7876
0.7794
0.7783
0.7780
0.7758

SF:

INTEN
4.
34.
3.

6.

9.

3.
38.
228.
6.
13.
40.
10.
7.
257.
4.

5.
11.
183.
278.
2.

4.
398.
2.

9.

4.
14.
15.
122.
9.
425.
6.
223.
152.
8.
181.
8.
8.
225.
22.
16.
191.
43.

1,111,1111:
9,999,9999:
F*F a m x [f

m x 1:
n x 9:
INTEN:
THETA:

L-P:

MULTI:

F*F
intensity

in degrees
Lorentz-polarization factor
multi-plicity factor

L-P
3.455
3.475
3.589
3.589
3.616
3.646
3.768
3.797
3.834
3.904
3.984
4.022
4.147
4.185
4.227
4.280
4.374
4.420
4.475
4.626
4.640
4.756
4.758
4.868
5.011
5.237
5.379
5.644
6.436
6.443
6.557
6.723
6.856
7.991
7.998
8.006
8.489
9.786

14.477
15.729
16.109
20.597

structure factor,F
superlattice reflections
fundamental reflection

Ni,Cu) "

n x {3 (Ni,Cu)

MULTI

beenbooooooooooooboooobooooooookbooooboobooboooobooboobooboooobookk

SF
111

1111
111

999
1111
1111

111
1111
999

111
111
9999
999
9999
1111

111
111

111
999

\O
\O
\OH

r—o
p—A
\OHHQr—HQHQ

t—o
H
D-l
H

f 1
.fig

11

THETA
61.01
61.16
62.01
62.02
62.21
62.42
63.23
63.41
63.64
64.06
64.52
64.73
65.40
65.60
65.81
66.08
66.52
66.74
66.99
67.64
67.70
68.18
68.18
68.62
69.15
69.94
70.41
71.22
73.30
73.31
73.58
73.94
74.22
76.29
76.30
76.31
77.03
78.65
82.20
82.81
82.97
84.48

46

structure, as described below.

The high temperature phase of the Ni-xCu-ZSSn (x-O-
30) is reported to have a D03 structure [6-9]. To
ascertain whether the high temperature phase is indeed the
D03 structure, the following experiments were performed:

(1) Specimens of the Ni-xCu-ZSSn (x: ranging from 10
to 22 atZ) alloys were encapsulated in evacuated quartz
capsules, kept for 1 hr at 1000°C and then quenched in ice
water. The quartz capsules were crushed immediately upon
entering the water. Examination of optical micrographs of
those quenched specimens did not reveal any evidence of
transformation during quenching and the high temperature
phase was retained unchanged. It should be mentioned that
the D03 phase was retained only for Ni3Sn alloys
containing high amounts of copper. In the case of the Ni-
14Cu-258n alloy, no D03 phase was retained. Instead, a
martensite transformation was observed. The details will
be described in section 4.4.

(2) A powder specimen of the Ni-17.SCu-253n alloy was
made and encapsulated in an evacuated quartz capsule.
After annealing at 1000°C, this powder specimen was
quenched in ice water without breaking the quartz capsule.
The crystal structune of the high temperature phase
retained in the quenched, powdered sample was identified
by means of X-ray diffractometry, using monochromatic Cu

K-alpha radiation. Angles of reflection peaks, measured

47

and theoretically calculated, are shown in Table 4.2,
along with their intensities. As seen in Table 4.2, all
the measured reflection angles and their relevant
intensities could be consistently explained by assuming a
D03 structure. Thus, it is certain that the high
temperature phase of the Ni-l7.5Cu-25$n alloy has the D03
structure. The lattice parameter of the structure was
determined to be a-5.910A. Murakami et al also measured a
lattice parameter of the D03 phase for (Ni,Cu)38n as a
function of copper composition by X-ray experiments.
Their reported data are shown in Figure 4.7. The lattice
parameter obtained in the present study is also shown in
Figure 4.7 as a large solid circle. As can be seen, the
obtained lattice parameter lies on an extrapolated curve
of those reported by Murakami et al [7].

Comparing Table 4.2 with Table 4.1, one can index
the unidentified reflection peak, 6, Table 4.1 as the
[220] reflection of the D03 structure. This reflection
has the strongest intensity in the D03 structure, and
therefore, it is possible that a small amount of the high
temperature phase might be retained which coexists with
the 2H structure in furnace-cooled specimens. However, we
still have peak 5, which cannot be identified. Table 4.3
shows reflection angles and intensities calculated by
assuming that the 0019 phase is present. As seen from the

table, some peaks cannot be indexed by the D019 structure

.0
01
co
1‘

.0
U1
(.0
N

O
U"
(D
O

Lattice parameter (nm)

Figure 4.7

48

 

 

— —i
I— l” -i
/’
[I
4“.
I
— _
l l I l

 

 

I
20 22 24

CU (at 0’0)

Composition dependence of the lattice
parameter of D0 structure in aa-quenched
Ni-xCu-ZSSn alloys.

16 18

49

Table 4.2 Angles, d-spacings, and intensities of reflection
peaks for a D03 structure of a Ni-17.SCu-258n
alloy. Lattice parameter used: a-5.910A;
cu‘kalpha 1.54439A; Debye-Scherrer method
used.

H K L D-SPAC THETA INTEN L-P MULTI SF
1 1 1 3.4121 13.08 182. 36.195 8 1111
O 0 2 2.9550 15.15 97. 26.481 6 1111
0 2 2 2.0895 21.69 2123. 12.043 12 9999
1 1 3 1.7819 25.68 90. 8.213 24 1111
2 2 2 1.7061 26.91 26. 7.382 8 1111
0 0 4 1.4775 31.51 325. 5.178 6 9999
l 3 3 1.3558 34.72 70. 4.213 48 1111
0 2 4 1.3215 35.76 32. 3.972 24 1111
2 2 4 1.2064 39.80 647. 3.280 24 9999
1 l 5 1.1374 42.76 21. 2.973 24 1111
3 3 3 1.1374 42.76 7. 2.973 8 1111
0 4 4 1.0448 47.66 229. 2.741 12 9999
1 3 5 0.9990 50.62 34. 2.738 48 1111
2 4 4 0.9850 51.62 34. 2.759 48 1111
0 0 6 0.9850 51.62 4. 2.759 6 1111
0 2 6 0.9345 55.73 429. 2.948 24 9999
3 3 5 0.9013 58.96 18. 3.221 24 1111
2 2 6 0.8910 60.08 19. 3.342 24 1111
4 4 4 0.8530 64.85 173. 4.044 8 9999
1 5 5 0.8276 68.92 51. 4.949 48 1111
1 1 7 0.8276 68.92 25. 4.949 24 1111
0 4 6 0.8196 70.42 27. 5.384 24 1111
2 4 6 0.7898 77.89 2098. 9.137 48 9999

1111:
9999:

superlattice reflection - 16(fA-fB)§
fundamental reflection - 16(3fA+fB)

Table 4.3

:1:

wwNNJ-‘HJ-‘kwHquUNJ-‘wauwHNWNNHOUDLQNNHNHNOHNHHQH

NI-IONn—H-H-IONOONOHHONOOHOo—NHNHOOOOOHHHOOOHOHOOO N
ww4>wH#ONHJ-‘LAOJ-‘Nwt-iuour-INuHOONuwHOMHNONHNh-OOHHO r

50

Angles and intensities of reflection peaks
for a D053 structure of a furnace-cooled
S

Ni-l4Cu-

n alloy.

for theoretical calculation:
c-4.260A; Cu-k alpha:
Debye-Scherrer method used.

D-SPAC

4.5951
4.2600
3.1240
2.6530
2.2976
2.2520
2.1300
2.0222
1.9325
1.7368
1.6609
1.6083
1.5620
1.5317
1.4414
1.4200
1.3567
1.3460
1.3265
1.2745
1.2665
1.2519
1.2436
1.2210
1.2079
1.1488
1.1260
1.1092
1.0993
1.0936
1.0650
1.0542
1.0413
1.0375
1.0233
1.0111
1.0027
0.9883
0.9761
0.9693
0.9662
0.9485
0.9448

INTEN

l7.
0.
42.
5.
89.
17.
94.
360.
4.
10.
2.

U!
C
O

mwuHHbouoouo

\IH

1" 0‘
“\OH
0 a a

H
O

H

HJ>
HumouHkaI-‘OH

L-P

68.253
58.263
30.069
20.965
15.118
14.433
12.674
11.210
10.057
7.760
6.952
6.421
5.974
5.692
4.904
4.730
4.244
4.167
4.029
3.686
3.636
3.549
3.500
3.373
3.304
3.029
2.942
2.885
2.856
2.840
2.774
2.756
2.740
2.736
2.728
2.729
2.733
2.749
2.772
2.789
2.798
2.863
2.880

Lattice parameters used

a=b-5.306,
1.544390A;

MULTI SF THETA
6 1 9.65
2 0 10.41

12 111 14.27
6 1 16.87
6 9 19.58

12 111 20.00
2 9999 21.20

12 999 22.39

12 1 23.49

12 1111 26.32

12 1 27.63

24 0 28.61

12 9 29.54
6 1111 30.19

12 0 32.30
2 0 32.85

12 111 34.59

24 1111 34.90
6 9 35.49

12 1 37.18

12 999 37.45

12 111 37.97

12 1111 38.27

24 111 39.11

12 999 39.62
6 9 42.10

12 9 43.16

12 999 43.98

24 0 44.48

24 1 44.77
2 9999 46.32

12 1 46.94

12 0 47.70

12 1 47.94

24 111 48.82

12 9 49.62

12 1 50.19

12 1 51.20

24 111 52.10

12 999 52.62

12 9 52.86

24 111 54.30

24 1 54.61

M#N§Mwa#MNHMW#CW#wwa#NU‘I
OOONt—IHWHl—‘HNOONNOWNOWOOHHO
mnFLflNHMNJ>WO§U|NWHLnHO#OUHN&O

0.9190
0.9079
0.9072
0.8984
0.8931
0.8843
0.8744
0.8684
0.8659
0.8520
0.8509
0.8464
0.8438
0.8377
0.8305
0.8253
0.8191
0.8172
0.8167
0.8112
0.8102
0.8041
0.7988
0.7810
0.7715

51

§

Ch
owwwamNHJ-‘OOOONHNNh-‘bo

190.
77.
41.
16.

3.038
3.132
3.138
3.228
3.288
3.401
3.553
3.658
3.706
4.016
4.045
4.167
4.245
4.445
4.725
4.959
5.290
5.403
5.433
5.819
5.893
6.443
7.057
11.786
34.941

1
1111

111
999
1111

1111
9999

111
111

1111
111

1111
111

9999
999

111

56.94
58.04
58.10
59.02
59.59
60.57
61.75
62.50
62.82
64.70
64.85
65.50
65.90
66.85
68.05
68.95
70.11
70.48
70.58
71.72
71.93
73.31
74.63
80.49
86.72

52

either. This fact suggests that there is a new phase. As
will be shown in the next chapter, the new phase can be
identified by precise analysis of electron diffraction
patterns. As commonly accepted, electron diffraction is
quite powerful in determining complicated crystal
structures.

Figures 4.8 and 4.9 show X-ray diffraction patterns
obtained from furnace-cooled powder specimens of the Ni-
xCu-ZSSn (x-14 and 20) alloys. Some of these reflection
peaks again could not be indexed by assigning any of the
following structures: D03, 2H, or D019. Those peaks all
correspond to the afore-mentioned unidentified reflection
peaks, although they are not consistent regarding
reflection intensity.

Figures 4.6, 4.8, and 4.9 suggest a possibility of
the existence of a new phase. These three figures all seem
to contain the (220)D03 reflection peak, although the
reflection angle of the peak is slightly different.
Intensities for the peak, however, appear to increase with
increasing Cu content, suggesting that the volume fraction
of the retained D03 structure increases with increasing
copper content. (Intensity calculations for reflection
lines of the 2H, D03 and D019 structures shown in'Tables
4.1-4.3 were carried out by using 8 Dec 20 computer.)

Table 4.4 shows the results obtained by Debye-Scherrer

method for the furnace cooled Ni-xCu-253n(x-l4 and

53

6lNi-14Cu-25Sn

 

Angle ( 2 6 )

Figure 4.8 Partial diffractometer pattern of a furnace-
cooled Ni-14Cu—258n alloy.

54

55N1-200u-258n

 

us us in: LB 42 141' .140 39 38
Angle ( 2 6 )

Figure 4.9 Partial diffractometer pattern of a furnace-
cooled Ni—200u-25Sn alloy.

Table 4.4
peaks hkl
1 110
2 011
3 101
4 020
111
5 x
6 120
7 200
8 002
9 x
10 121
11 201
? x
12 211?
? x
13 122
14 202
15 212
16 013
17 040
18 320
19 123
20 203
21 240
22 042
23 322
241
24 401
25 004
26 242
27 402
28 124
204
29 x
30 243
? x
31 403

160

55

Debye-Scherrer data obtained from the alloys,
Cu-k alpha used.

furnace-cooled from 1000°C.

(cal) (meas)

8
11
7
18
5

?
56
27
103
?
289
141
?

7

?
66
32
7

A
207
100
136

68
22
87
174
130
65
43
45
23
48
24

183

99
34

Ni-14Cu-253n
intensity

V
w
1’"
W

<
taawtwmmmmaat

<

<

<
S

aatmmma

stat:

Qt '95 t

theta(degree)
(cal) (meas)
12.91 13.07
13.30 13.45
14.39 14.75
16.65 16.95
16.67
? 19.00
19.51 19.55
20.06 20.15
21.09 21.14
7 21.86
22.29 22.42
22.79 22.87
? ?
24.39 23.99
? ?
29.40 29.47
29.81 29.97
31.16 31.37
33.95 33.94
34.97 35.07
36.08 36.10
39.40 39.34
39.76 39.83
41.91 41.82
42.59 42.62
43.64 43.57
43.77
45.17 45.08
46.03 46.16
49.35 49.34
50.77 50.95
52.50 52.64
52.86
? 55.98
59.10 58.99
7 ?
60.71 60.48
61.16 ?

Ni-17.SCu-253n

inten

theta

peaks (meas) (meas)

1

CDVO‘UIkUJN

11
12

13
14
15
16
17
10
19

20
21
22

23
24
25
26

VW

tilts

mttmmaa

as:

555‘

13.10

19.50
20.13
21.15
21.77
22.40
22.90
23.07

25.70
29.28
29.85
31.45

34.92
36.04
39.24
39.88
41.77
42.39
43.55

45.20
46.00
49.16

55.77
58.86
59.86
60.31

peaks

32

33

34
35
36

37
38
39

40
41

42
43
44
45

hkl

413
304
432
440
161
351
520
015
441
105
423
044
521
324
162
125
205
442

x
522

x
244

(cal) (meas)

3

6

9
38
227
13
41
7
257
6
11
183
279
400
121
429
226
153

183

229

56

Ni-14Cu-258n
intensity

W

t

(ta-0's)

ttttaaafltat

<

theta(degree)
(cal) (meas)
62.11 62.31
62.12
62.28
63.28 63.44
63.41
64.09 ?
64.61 ?
65.53 65.27
65.66 65.51
66.21 66.51
66.63
66.82 66.83
67.09 67.79
68.30 68.26
71.24 ?
73.49 73.00
74.14 73.65
74.33 .
? 75.73
76.49 76.33
? 78.17
78.83 78.81

Ni-17.5Cu-258n

inten

theta

peaks (meas) (meas)

27
28

29

30
31

32
33

35
36
37
38

39
40

t m

to 03535

(1253

63.17
63.51

64.59

65.33
66.68

66.80
67.81
68.10
72.94

73.54

75.74
76.63
77.98
78.41

57

17.5atZ) alloys. Measured intensities are denoted by the
following indications: vw(very weak), w(weak), m(medium),
s(strong) and vs(very strong). From these data, a new
phase is known to coexist with the 2H and 003 phases in
(Ni,Cu)3Sn alloys when they are furnace-cooled. Most of
the observed reflection lines can be indexed by assigning
the 2H structure, while some of them cannot be indexed
even by assigning the D03, 2H, and D019 structures.

The X-ray diffraction data obtained by both the
Debye-Scherrer method and the diffractometer method were
in very good agreement each other; However, the results
obtained in the present study are inconsistent with the
reports by Murakami et al [7—9], which clearly stated
that, in the range of copper contents between 8 and about
18 atZ, the 2H phase is the only one present at

temperatures below 700°C.

58

4.3 Microstructures in Furnace—cooled Specimens

As shown in previous sections, X—ray diffraction data
and optical micrographs clearly indicate that there is
another phase in addition to the 2H phase in furnace-
cooled specimens of the three Ni—xCu-ZSSn alloys (x-14,
17.5, and 20). In this section the crystal structure of
this proposed, new phase will be identified by means of
electron. diffraction. Also, results of detailed
examinations of microstructures for furnace-cooled
specimens will be shown.

Disks of the Ni-xCu-2SSn alloys (x=14 and 20) were
encapsulated in evacuated quartz capsules and annealed at
1000°C, where the high—temperature D03 phase is known to
be stable. The disks were kept at this temperature for
about 1 hr, then furnace-cooled.

The transmission electron microscopes primarily used
in this study were Hitachi HU—200F and H800 microscopes,
which were operated at 200 keV. A Kratos EM 1500
microscope located at National Center for Electron
Microscopy, Berkeley, CA was also used for some work.

This microscope was operated at 1500 keV.

59

443.1 Microstructures of a Furnace-cooled Ni-14Cu-258n

Alloy

Figure 4.10(a) is an electron micrograph, showing a
typical microstructure of the furnace—cooled alloy. In
the figure, two distinct regions are seen. One is a
region composed of bands. This region is surrounded by an
irregularly curved boundary. These bands are seen to
contain densely-packed internal structures, whose
contrastsresembleones for either thin twinscnrstacking
faults. The bands are bounded by slightly dislocated
areas, which have the same crystal orientation. .This can
be known from the continuation of bend contours. Because
bands vary in size and orientation, regions containing
bands are very complicated. To clarify relations further
between the bands and the associated areas of
dislocations, a part of Figure 4.10 was enlarged.

The enlarged micrograph is shown by the single-headed
arrow in Figure 4410(a). A brief survey of this area leads
one to the incorrect idea that three bands form one spear,
as often reported for Cu-Al-Ni alloys [32]. In the case
of the Cu-Al-Ni alloy, the spear consists of two 2H
martensite variants. The side bands have twin-like
internal structures, while the center one has no such
internal structure, but contains a dislocated structure.

It is clear then that this spear-like structure is

Figure

4.10

60

Microstructure of a Ni-14Cu-258n specimen,
furnace-cooled from 1000°C. (a) bright
field image of 2H and d2H phases. (b)
diffraction pattern, taken from the top
left-hand side of a 2H region: [111] zone
axis. (c) diffraction patterns, taken frgm
the central region of the micrograph: [111]
zone axis for 2H and d2H. (d) angles
between two reflection spots.

 

 

(d) angle(degrees)
2H d2H

Si‘gj obs.
011.;10 66.5 72.3 72
911.101 55.5 52.0 54
110-101 58.0 55.8 54

 

 

 

 

 

61

different from those observed in the Cu-Ni-Al alloy. It
can be seen in Figure 4.11 that the bands with twin-like
internal structures and areas with dislocations are
alternatively formed.

The other region shows a slightly different
microstructure that contains a low density of
dislocations. The morphology of this region is
essentially the same as that of dislocated areas shown
above. Figure 4.12 is an enlarged micrograph of an area
pointed by the double-headed arrow in Figure 4.10(a). It
can be seen in Figure 4.14 that a sub-boundary was
constructed by dislocation networks. In the area enclosed
by the networks, one can see typical fringe contrasts of
stacking faults. These stacking faults are considered to
have been formed by dissociation of boundary
dislocations.

Figure 4.13 is 21 high magnification electron
micrograph of another furnace-cooled Ni—l4Cu-258n
specimen. Diffraction patterns were taken from a large,
bright area in the upper right-hand side of the
micrograph. It was found that the crystal structure of
the area is a 2H structure. There is a low-angle
boundary, consisting of at least two sets of dislocation
arrays which run from the top left-hand side to the bottom
right-hand side of the micrograph. In bright areas of the

2H phase, dislocation loops and vacancy loops are seen.

62

Figure 4.11 Enlarged micrograph of an area indicated
by the single-headed arrow in Figure 4.10.

 

63

Figure 4.12 Enlarged micrograph of an area indicated
by the double-headed arrow in Figure 4.10.
Note fringes of stacking faults and
dislocation networks in a 2H region.

 

64

Figure 4.13 High magnification micrograph of a furnace-
cooled Ni-14Cu-253n specimen, showing many
microstructural features.

 

65

These loops are thought to be associated with the
formation of the fine bands seen in Figure 4.13. These
fine bands were found to be a new phase having a distorted
2H structure, which is slightly different from that of the
normal 2H phase. Identification of this new phse will be
described in section 4JL3. It should be noted that the
distorted 2H phase has not been observed to cut across
low-angle boundaries.

Two noticeable microstructural features are seen in
Figure 4.13. One is the band denoted by A; the other is
in the areas denoted by B,C,D, and E. The former
microstructure clearly reveals characteristics similar to
those of the bands formed in furnace-cooled specimens.
Close inspection of band A discloses that the band is
thick and contains many planar faults. These planar
faults were probably formed when bands nucleated in
regions of the D03 phase. From the appearance of the
bands seen in Figure 4.13, it is obvious that the bands of
distorted 2H phase grew to hold orientation relationships
with its D03 parent phase. From band A, one knows that
the other bands seen in Figure 4.13 are similar but their
width is not so large as those for marked regions 'a' and
'b' in band A. Therefore, the contrast of those bands
appears complicated.

The other microstructural feature in Figure 4.13 is

one that demonstrates the nucleation process of bands.

66

Close observation of areas 0 and D in particular shows
that each area contains striations oriented :hi two
directions. Either of the directions is parallel to each
other in the two areas. Microstructures in areas C, D,
and E are otherwise very similar to those observed in
nuclei of a martensite in a Aqu alloy studied by
Mukherjee et al [16]. It should be noted that these
striations are not parallel to the stacking faults
observed in bands A.

In Figure 4.13, it can be seen that the upper right-
hand side of the figure does not contain any band
structures, while the lower part contains many bands of
varying size that exhibit stacking faults. These two
areas are divided by dislocation networks that create a
low-angle boundary. Because the areas have the same 2H
structure, and their crystal orientation is almost the
same, it is necessary to explain their microstructural
differences. The explanation can be made by considering
the following phase decomposition that may occur during

furnace-cooling of (Ni,Cu)3Sn alloys:
D03 ---> 2H(l) + D03 ---> 2H(l) + (2H(2) + d2H)
where d2H is the distorted 2H phase. Areas of the 2H

structure transformed at high temperatures from the 003

phase are denoted as 2H(l), and areas later transformed to

67

the 2H structure at lower temperatures is denoted as
2H(2). Their lattice parameters differ slightly from each
other, both because of diffusion processes and because the
phase boundary involved is probably parallel to the
composition axis of the composition-temperature phase
diagram. The dislocation networks seen in Figure 4.13
are, therefore, thought to have been formed to accommodate
lattice misfit between those two 2H regions. In other
words, low angle boundaries are probably original phase
boundaries between 2H(l) and D03 regions.

Figure 4.14 (a) is an electron micrograph of a
furnace-cooled Ni-l4Cu—258n specimen. Because this figure
is a low magnification photograph, macrostructural
differences are clearly seen between a two-phase region
with many bands and regions with dislocations.
Diffraction patterns taken from a: lower region with
dislocations and a central region with bands corresponding
to Figure 4.14(b) and (c), demonstrate that the latter
banded region is composed of two phases, the 2H phase and
the new distorted 2H phase. More importantly, the 2H
phase in the central two—phase region has nearly the same
crystal orientation as that of the lower region with
dislocations; the foil normals of those regions are
analysed to be the [234] zone axis. This result is
consistent with the observations of dislocation networks

seen in Figure 4.13.

68

Figure 4.14 Microstructure of a Ni—l4Cu-253n specimen,
furnace-cooled from 1000°C. (a) bright
field image of 2H and d2H phases.

(b) diffraction pattern of [234] zone axis
taken from a 2H region. (c) diffraction
patterns taken from a two-phase region.

 

 

69

It should be noted that the diffraction patterns taken
from the central region contain two other zone axes,
marked as dotted lines in Figure 4.14(c). Inspection of
the diffraction patterns for these zone axes demonstrate
that they are for two variants that appear as bands in the
central region. The crystal structure of the variants
will be identified in section 4.3.3.2.

Figure 4.15(a) is another micrograph taken from the
same specimen as that used for Figure 4.14. A two-phase
area with variants is seen to beienclosed by dislocation
networks in which surface etching is evident. 2H regions
are also seen within the enclosed area. Figure 4.15(b) is
a diffraction pattern, taken from a region within the
enclosed area having bands. It was found that streaks in
Figure 4.15(b) are all perpendicular to fine
microstructures in bands.

Figures 4»16(aLKb), and (c) are high magnification
micrographs, taken from the framed areas, A, B, and C, in
Figure 4.15(a), which show detailed microstructures of
the low-angle boundaries. These boundaries are zigzag and,
as was seen in Figures 4.10 and 4.14, no bands cross these
zigzag low-angle boundaries. Striations clearly seen in
Figures 4.16(a), (b), and (c) are parallel to each other,
and are perpendicular to streaks in the diffraction
pattern in Figure 4.15(b). These striations are stacking

faults that were formed as lattice invariant strain. This

70

Figure 4.15 Microstructure of a Ni-14Cu-258n specimen,
furnace—cooled from 1000°C.(a) bright field
image of 2H and d2H phases. (b) diffraction
pattern taken from an area containing bands.

 

 

71

Figure 4.16 High magnification micrographs taken,
(a) from area A; (b) from area B; and
(c) from area 0 in Figure 4.15.

 

72

Figure 4.16 High magnification micrographs taken,
(b) from area B in Figure 4.15.

 

73

Figure 4.16 High magnification micrographs taken,
(c) from area C in Figure 4.15.

 

74

invariant strain is necessary for the new phase to
nucleate and grow in the D03 matrix. Figure 4.16(a) shows
a moire fringe pattern, indicating complication of this
microstructure: the nature of the moire fringe pattern is
not clarified.

In Figures 4.10 to 4.16, focus was placed on
microstructural features of the new phase, formed as
bands in the furnace-cooled Ni-14Cu-258n. Note, however,
that dislocated areas of 2H-phase are predominant. A
typical example of a 2H-phase area which demonstrates
dislocations and thin planar faults is shown in Figure
4.17. Trace analysis revealed that the thin planar faults
were on (101) planes, one of the two twin planes that were

reported for 2H martensite by Pak[6] and Ostuka[32].

75

Figure 4.17 Electron micrograph of a 2H region taken from
a Ni-l4Cu-258n specimen, furnace-cooled from
1000°C. Note (101) planar faults.

 

 

76

4JL2 Microstructural Morphology of a Furnace—cooled

Ni-ZOCu-ZSSn Alloy

Specimens of this alloy were annealed at 1000°C and
furnace-cooled. Figure 4.18 is a typical example of
microstructures obtained in the specimens. It can be seen
from this figure that bands contain internal faults and
occur in a zigzag form. Areas bounded by these bands are
heavily bent by buckling in thin areas possibly caused by
the elastic strain field between a band and the neighbor
area.

Figure 4.18 also shows many dislocations at or near
low-angle boundaries, in which the new phase nucleated or
by which the new phase was blocked to grow; these
dislocations are considered tn) be transformation
dislocations generated to accommodate transformation
strain. A narrow band was probably formed after wide
bands were formed at higher temperatures.

The micrograph shown in Figure 4.19 demonstrates that
the microstructural morphology in the right-hand side of
this figure is similar to that observed in Figure 4.18.
Bands in Figure 4.19 also exhibit a zigzag form. Bands
are known to have a certain width from the number of
boundary fringes. It is interesting that both dislocation
lines and loops are seen, marked by the single-headed

arrows. More important is the microstructure seen in the

Figure

4.18

77

Bright field image of a Ni-ZOCu-ZSSn specimen,
furnace-cooled from 1000°C to room
temperature. Note dislocations and planar
faults in 2H regions and planar faults in d2H
regions.

 

 

 

 

:92 2.5mm
.\ _ . 5...

 

78

Figure 4.19 Bright field image of a Ni-ZOCu-ZSSn specimen,
furnace~cooled from 1000°C, showing zigzag
bands.

 

 

 

79

left hand-side of the figure. Here, wide bands coexist
with uniformly distributed internal faults. These wide
bands formed alternately with dislocated areas. In the
dislocated areas, there are small-sized 2H bands showing
no internal faults. Some dislocations are also found in
internally faulted bands, marked by the double-headed
arrow.

A relation between wide bands with fine internal
faults and narrow, zigzag bands within dislocated areas is
shown in Figure 4.20. From this figure and optical
micrographs taken from furnace-cooled specimens, one can
see that the wide bands correspond to the rod-shaped
structure containing fine twins, while the zigzag bands
within dislocated areas correspond to areas surrounded by
the rod-shaped structures.

Differences in microstructural features between the
two Ni—xCu-ZSSn (x - 14 and 20) alloys are summarized as
follows: (1) the volume fraction of the 2H phase to the
new phase is lower in the Ni-ZOCu-ZSSn alloy than in the
Ni—14Cu-258n alloy, suggesting the location of phase
boundaries for these two phases, and (2) the 2H phase has
a larger number of (101) planar faults in the Ni-ZOCu-ZSSn
alloy, while the 2H phase has a fewer faults in the Ni—
14Cu-253n alloy, suggesting a relative change in lattice
parameters of the 2H and the new phase as a function of

copper content. No detailed observationscflfthe lattice

80

Figure 4.20 Bright field image of a Ni-ZOCu-ZSSn specimen,
furnace-cooled from 1000°C to room temperature.
Note many dislocation loops and dislocation
lines in 2H regions.

 

 

 

81

parameter changes were made.
4.3J3 Identification of Phases by Electron Diffraction

Electron diffraction was used to investigate the
crystal structures of both the banded and the slightly
dislocated regions which were formed in the Ni-14Cu-258n

alloy that was furnace-cooled from 1000°C.
4JL3.1 Crystal Structure of Slightly Dislocated Regions

Figure 4.21 shows four zone axes of selected-area
diffraction patterns taken from slightly dislocated
regions. It is seen that all the diffraction patterns can
be indexed by assigning the 2H structure (gamma-Cu3Ti-
type) which has lattice parameters of a=4.495A, b=5.380A,
and c-4.285A, as measured in X-ray experiments. The four
zone axes in Figure 4.21 are known to be [100], [2T0],
[0T1], and [0T2]. The slightly dislocated regions are

found to be the 2H phase.
4JL3.2 Crystal Structure of Banded Regions
The crystal structure of the banded regions (a new

phase) was also identified. Figure 4.22 shows three

selected-area diffraction patterns taken from banded

82

Figure 4.21 Electron diffraction patterns taken from
2H areas of Ni-l4Cu-25Sn specimens furnace—
cooled from 1000°C. Zone axes: (a) [100]:
(b) [210]; (c) [0T1]; and (d) [oIz].

 

i Tum

up,

 

83

Figure 4.22 Electron diffraction patterns taken from
d2H areas of Ni-14Cu-258n specimens furnace-
cooled from 1000°C. Zone axes:_(a) [001]:
(b) [010]; and (c) [001] and [214].

 

84

regions. These three diffraction patterns contain a-, b-,
and c-axes; these three axes are not perpendicular to each
other. From the three diffraction patterns, the three
angular lattice parameters can be measured and are as
follows: alpha - 85°, beta - 86°, and gamma 2 84°.

By adopting these angular lattice parameters and the
unit cell lengths estimated as a=4.53A, b-5.31A, and
Ca4.34A, one can index all the diffraction patterns shown
in Figures 4.22(a,b,c) under the assumption that the
atomic arrangement of the new phase is the same as that of
the 2H phase. Tables 4.5 and 4.6 show the calculated d-
spacings and angles used for indexing diffraction patterns
in Figure 4.22. The table also includes the d-spacings and
angles measured from Figure 4.22. The calculated values
are seen to be in good agreement with the measured ones.
It should be mentioned that the crystal structure of the
new phase identified here can consistently explain the
unidentified X-ray peaks in Figures 4.6. 4.8 and 4.9. The

crystal structure identified is denoted hereafter as d2H.

85

Table 4.5 D-spacings of the d2H phase

hkl d-spacing
010 5.25
100 4.46
001 4.27
110 3.62
011 3.45
101 3.18
020 2.62
111 2.92
120 2.39
021 2.32
200 2.23
121 2.19
210 2.15
002 2.14
012 2.04
201 2.03
211 1.99
102 1.97
112 1.93

203 1.22

86

Table 4.6 Angles between two planes for d2H
and 2H structures

(h1k111)l(h2k212) observed 02H 2H
angles
Figure 4.22(a) 010.100 84 84 90
[001] 010.110 46.5 45.6 50
zone _
100.120 55 54.4 59.1
Figure 4.22(b) 001.100 85 86 90
[010] 201-100 26 26.8 27.7
zone
001.101 42 42.1 43.6
Figure 4.22(c) 120.201 70 70.4 63
[212] 201.321 32 31.8 28.9

zone

87

4.4 Microstructure and the Crystal Structure of As-

Quenched (Ni,Cu)3Sn Alloys

Several (Ni,Cu)3Sn alloys were quenched from 1000°C
to ice water in an attempt to retain the high-temperature
D03 phase and its product phase. However, even with a very
rapid quenching by crushing the encapsulated quartz tubes
in ice water, no dual phases could be retained. For
alloys containing Cu ranging from 16 to 22 atZ, the high-
temperature D03 phase was retained, whereas for an alloy
with a copper content of 14 at! Cu, a martensite was
formed as a product phase of the D03 phase. These
observations disagree with data reported by Murakami et a1
[7-9].

Figure 4.23 is an optical micrograph of the quenched
Ni-léCu-ZSSn alloy, showing a plate-like martensite. This
martensite reveals fine striations, indicating fine planar
structures. Bright and dark bands are martensite
variants, and no parent D03 phase is seen. The
microstructure of the martensite bears resemblance to that
of the thermoelastic martensite in a Cu-Ni-Al alloy[32].
However, the martensite formed in the Ni—léCu-ZSSn alloy
did not show any thermoelastic characteristics. Detailed
microstructures of the martensite are seen in transmission
electron micrographs.

Figure 4.24 is an electron micrograph of a martensite

88

Figure 4.23 Optical micrograph of a Ni-thu-ZSSn

specimen, quenched from 1000°C to ice

water, showing a plate-like martensite
formed from a D03 phase.

 

89

Figure 4.24 Electron micrograph of a 2H martensite
(denoted as ZH') formed in a Ni-léCu-ZSSn
specimen, quenched from 1000°C to ice water.
Note a high density of (121) twins.

 

90

with a high density of microtwins. Trace analysis of the
micrograph showed that the twinning plane was the(121)
plane of the 2H structure. As will be shown below, the
crystal structure of the martensite was identified to be
the 2H structure.

A similar microstructure is shown in Figure 4.25(a),
where two martensite variants are present. The
diffraction pattern taken from the area indicated by the
single—headed arrow in the figure is shown in Figure
4.25(b). Figure 4w25 (c) is taken from the area by the
double-headed arrow. Both can be indexed as [l20] and
[124] by assigning them the 2H structure. Several other
diffraction patterns were also obtained from the
martensites and all could be indexed by the 2H structure.
To obtain further details of the martensite micro-
structure, diffraction patterns were taken from a
martensite variant located in the upper portion of the
micrograph. Figure 4.25(c) represents the diffraction
patterns thus obtained, which clearly exhibit a twinning
relationship. Twin planes in Figure 4»25(a) were found to
be the (121)2H twins, as observed in Figure 4.24. Some
(121) twins in the present alloy contain fine striations,

which are stacking faults.

Figure

4.25

91

Microstructure of an as-quenched Ni-léCu-ZSSn
specimen, showing two martensite variants
containing twins. (a) bright field. (b)
diffraction pattern of [120] zone of 2H,
taken from an area indicated by the single-
hgaded arrow. (c) diffraction pattern of
[124] zone, taken from an area indicated by
the double-headed arrow.

 

O

42%.
gm .

 

4.5 Aging Effects of the Ni-20Cu-258n Alloy

4.5.1 Results

The X—ray and electron diffraction analyses have
revealed that either the two phases, 2H and d2H, or the
three phases, 2H, d2H, and D03, are formed in Ni-xCu-ZSSn
alloys (x ranging between 14 and 20 atZ) when they are
furnace-cooled from 1000°C. Because of this heat
treatment, however, I“) transformation temperatures from
the D03 phase to the 2H and d2H phases are known. To
obtain information about transformation temperatures,
aging studies were carried out. The alloy had a Cu content
of 20 atZ, and it was quenched from 1000°C to retain the
high temperature D03 phase. Samples were then aged for 24
hr at 650°, 600°, 500°, and 350°C. After these aging
treatments, the specimens were again quenched in ice water
and investigated by X-ray diffractometer.

In the case of the 650°C aging, no phase
transformation of the retained D03 phase was found to have
taken place. In the range of angles (theta in degree)
between 18 and 60, only three sharp reflection peaks were
observed at angles of 21.69, 25.57, and 47.37°. By
comparing with Table 4.2, these angles were indexed as the
fundamental reflections of the D03 structure, 022, 113,

and 044. Figure 4.26 shows the portion of X-ray

92

93

 

 

 

 

 

 

I.

   

..o.

 

 

 

 

 

 

 

 

. . . . .I.I
IILI1IIvlv- I; II. IIIJIIIII: “I. I..-
..... o . .
I... I. .
II....I .
I ..III.
I II A
I II , .
II III .1 CI
III III“. I I .I. s l
I III. III I
III II. r. . I. I
I .II a . .
I IIIII? I 1 11> . I
loIIII .1 o I I o

\—
III.II .. I .I I . .1
IIIII . I I I. 4 ........
IIIH 1:10 I I . I .II .II IIIII I II
III .5I II I I I . II III» I1. III...

1A
II III 0 I “ . I I.III11
III I IO- -.10 III.- .. a I
.I IIV‘ I. I 1 I1 . b
IIIII1QII II I II III a
Ir 1»
4

II: I. l I 1‘ I. III .
IIIOIIIIeg II . J . I. I
I- 0' n. . . - I l‘. .
IIr VIII II r.II In . .II 11. 1
III .IIIIIII II F
II II I II I. I I I .I I IIIII I N
I. .II.‘ II IIIIII
II .III ‘ l1 4 III .....
I. In . . .
I. .0 u r o .. II
I'I II. . I I‘ I . O .
III Io. - 1... I e I. I
I I g I . H .o . n .e . . ..

k1
III1I 1 III I . III VIII
tell. ”I III. . . III 0 III.
I11.IIIO III I: A .III. II
ITIIII s ...... I I 0_ ....... .I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I III I III
I‘II J
10" ’ I 'II ' III -+ I.
III I c I II . I .. I1
IIIIIII lull III .011 in. ...... Hui I
II. IIIQI III II. I n . I I .r 1 I
IIIcIII I I I . . II ..I I
10.1111.” ...... u .. n .14. N ..
I1QII , .1 I . ......qu II
I» I: e I I I . 6 III .1
114 III
I. III IOI; II 4.10 e I II.
II.III .J. I. 8 I “III 1 II
I .II .Ib' II I I. I I t I III
II.OI1II. .. I 1 o I I III
F
III. II .10 I . ¢ I1 .-
-IIOIII1uu. . .o ‘I .I
II...OII..OI . 1.! ... 4
IIIIIIaI o . I I1 IQ. III
III L L
I...‘ .I . 1&1. ml '1‘
II‘IIIOIIIIIIII $ oquI1 I ...
I .Q'. '6‘ I O
I .II n . . o 1 I o 11.
In all... I . . 1
-.IIIOIII Ian”. I1. I A
I, VII *.0 .m... I L
' 1.-.. I I. I II ‘ ‘ .1 II
L
I 4I .aI. J I a II I
II III . I . .o I IIIIII
-II ..IIIIO 1 I I
II III .I 1 1 . .
L
I II . 4 I
In III . I .
- l.‘ . I l I -
'II‘-‘ I I D .-
IN, I
-IIOII I II I o .
I III 0. I I .. .0 I.
.I.I. III I} III I I II II . I
I- ‘4'. II C ‘ h-
1! I u o
. .
r q . v.
,4
...

 

 

m1w9w

 

 

Partial diffractometer pattern of a Ni-ZOCu-

Figure 4.26

in

quenched in ice water from

1000°C aged 24 hr at 650°C, and quenched

2SSn alloy,
ice

water.

94

diffractogram obtained from a specimen aged at 650°C,
which presents the strongest 002 peak of the D03
structure. As seen in this figure, two broad peaks were
also observed, which appeared at angles of approximately
23.2 and 26.70. These peaks were found not to be indexed
by assigning them to any of the crystal structures already
known in this present study.

0n the other hand, in the case of the 500°C aging,
the phase transformation of the D03 into the 2H phase is
evident. The reflection peaks observed from an aged
specimen are listed in Table 4.7, which can all be indexed
by both the D03 and the 2H structures. In Figure 4.27, a
portion of X-ray data is shown, whose angle range is
corresponding to that for Figure 4.26. As seen in these
figures, their aging behavior is evidently different.

For a specimen aged at 600°C, essentially the same
transformation was observed as that obtained from the
specimen aged at 500°C: both the 2H and the D03 phases are
seen to coexist at 600°C (Table 4.8). Note that all
reflection peaks in the table can be indexed by assigning
them to the D03 and the 2H structures, although some
discrepancies are seen in the intensities of the peaks.
Comparing angles shown in Tables 4.7 and 4.8 with angles
listed in Table 4.1, it is also seen that the lattice
parameters of the 2H phase formed in the alloy with a Cu

content of 20atZ are the same as for the alloy with a Cu

Table 4.7

Number
peaks

11
12

of

95

X-ray reflection peaks observed
in a specimen aged 24hr at 500°C.

Angle(9°)
measured

15.12
19.50
19.93
21.16
21.63
22.29
22.88
26.80
31.38
34.92
39.56
42.45

Intensity
measured

VW

VW

vs(?)

VVS

Index
200(D03)
120(2H)
200(2H)
002(2H)
022(DO3)
121(2H)
201(2H)
222(DO3)
004(D03)
133(DO3)
224(D03)
115,333(D03)

Figure 4.27

96

 

Partial diffractometer pattern of a Ni-ZOCu-
25Sn alloy, quenched in ice water from

1000°C aged 24 hr at 500°C, and quenched in
ice water.

Table 4.8
Number of Angle(0°)
peaks measured
1 12.83
2 13.22
3 14.27
4 16.58
5 19.40
6 19.90
7 20.95
8 21.54
9 21.65
10 22.35
11 22.62
12 24.92
13 25.25
14 29.20
15 29.60
16 34.80
17 34.91
18 35.80
19 39.14
20 39.47

97

X-ray reflection peaks observed

in a specimen aged 24hr at 600°C.

Intensity
measured

W

W

VVS

VW

VW

Index
110(2H)
011(2H)
101(2H)
111,020(2H)
120(2H)
200(2H)
002(2H)
210(2H)
022(D03)
121(2H)
201(2H)
112(2H)
311(003)
122(2H)
202(2H)
040(2H)
133(003)
320(2H)
123(2n)

203(2H),224(DO3)

98

content of 14atZ.

Another aging experiment was also made at 350°C in an
attempt to observe the d2H phase. The X-ray data obtained
from this sample show the presence of the d2H phase, as
shown in Figure 4.28. All reflection peaks were very
broad, and the intensities are not strong. Only (121)d2H
and (022)D03 peaks are clearly identified. Also, electron
diffraction analysis of specimens aged at 350°C was
carried out, and the presence of the d2H phase was
revealed. An example of the diffraction patterns is shown

in Figure 4.29.

4.5.2 Discussion of Aging Results

With the 650°C aging, only a few sharp reflection
peaks were observed, in addition to two broad lines.
Because coarse-grained specimens were used for the X-ray
diffractometer experiment, the appearance of fewer peaks
may have been caused by preferential orientation of grains
that grew during the aging. The line broadening phenomena
are more interesting; the presence of the broad lines
indicates that some phase decomposition or some pre-
transformation might occur at 650°C prior to the phase
transformation from the D03 to the 2H phase. Transmission
electron microscopy and diffraction analysis of specimens

aged at 650°C for 20 hr support the idea that a spinodal

99

'5).

 

. . ’ ’
l ‘ Q—

Figure 4.28 Partial diffractometer pattern of a Ni—20Cu-

25Sn alloy, quenched in ice water from

1000°C, aged 24 hr at 350°C, and quenched in
ice water.

100

Figure 4.29 Diffraction pattern of a [214] zone, taken
from a d2H region in a Ni-20Cu-258n
specimen, aged at 350°C for 24 hr.

 

 

 

101

decomposition took place in the D03 phase. This
observation is discussed in further detail below.

A typical modulated structure obtained from the aged
specimen is shown in Figure 4.30(a), suggesting a spinodal
decomposition. A diffraction pattern taken from the same
area seen in Figure 4.30(a) is shown in Figure 4.30(b).
As is evident in this figure, a spinodal decomposition has
occurred. The following are characteristics of the
spinodal decomposition seen in Figure 4.30(b): (1) the
satellite reflections, e.g., at 220 and 400, are
obvious; and (2) their spacing is seen to be independent
of the distance from the origin, 000, as would be expected
for the spinodal decomposition of a given wavelength [33].
It should be noted from the appearance of the superlattice
reflections, e.g. 200 and 020, that D03 order is
preserved despite the decomposition. This is consistent
with the X-ray data shown in Table 4.2. Two more
characteristics of spinodal decomposition are seen in
Figure 4.30(b): (1) there are no satellites of the
superlattice reflections, e.g. 200 and 020; and (2) no
satellites of the transmission spot, 000. In Figure
4.30(a), the modulated structure is seen to have two
directions, which are perpendicular to the satellite
reflections in Figure 4.30(b).

From the previously discussed evidence, it is

apparent that the D03 phase in the (Ni,Cu)3Sn alloys is

Figure

4.30

102

Spinodal structure of a Ni-20Cu-253n
specimen, aged 24 hr at 600°C, following
ice-water quenching from 1000°C.

(a) modulated structures of a D03 phase.
(b) two sets of satellite reflections,
perpendicular to the modulated structures.
Note no satellite of superlattice
reflections. [100] zone.

 

 

 

103

present at 600°C but it has undergone the spinodal
decomposition. In this study, no further investigation
was made on the spinodal structure to obtain the final
structure of the decomposed phase. However, by taking
the results of a recent study of (Cu,Mn)3Al alloys [34]
into account, one can speculate upon the final structure.

The (Cu,Mn)3A1 alloys were studied in detail by
electron microscopy and diffraction analysis, tnul it was
found that the final structure of the spinodal
decomposition in these alloys was the dual structure, the
D03 structure of Cu3A1 and the L21 structure of CuZMnAl
[34]. Figure 4.31 shows phase relationship among M3Sn
(M-Ni, Cu, and Mn) compounds. As seen in this figure, Ni—
Cu-Sn alloys have both the D03 structure of Ni3Sn and the
L21 structure of CuzNiSn. From such structural
similarities between the present (Ni,Cu)3Sn alloys and the
(Cu,Mn)3Al alloys, it can be concluded that the final
structure of the decomposed D03 phase of the (Ni,Cu)3Sn
alloys would be the D03 structure of either Ni3Sn or

(Ni,Cu)33n and the L21 structure of CuzNiSn.

104

 

 

,” [35] 7 7 [29]
Mn3Sn,D019N12MnSn,L/21 (Ni,Cu)3Sn,D03 903 + L21
Schubert, 1946 Castelling,l953 Watanabe,1981 ?

[36] [37] , Cu‘25 atZ_ ’,
m ,z
4’
, _ Aa"'

N13Sn(L),D019 = N13Sn(H), D03 ~~
Rahlfs,l937 Schubert, 1956 ‘\\;::;::~
‘~ “

 

 

 

 

 

 

 

CUZMnSn(H),L21 CuZNiSn, L21
Carapella,l941 Klyucharev,l939
[38] [39]
CugSn, D03
Hendus,l956 M: Massive Transformation

 

[40] B: High Temperature Phase
L: Low Temperature Phase

Figure 4.31 Phase relationship among M3Sn (MaNi, Cu ,Mn)
compounds.

105

4.553 Conclusions Based on Aging Experiments

From the four aging experiments employed, the
following conclusions can be made.

(1) the high temperature D03 phase is stable above
650°C and below the melting point of the alloy (1022°C
measured by DTA); however, it is transformed in part to
the 2H phase between 500 and 650°C to be the dual phases,
the D03 and 2H phases.

(2) the D03 phase appears to be metastable at
temperatures of about 650°C. From the appearance of the
line broadening (Figure 4.28), the D03 phase is thought to
have a pre-transformation, as is often observed in many
ordered martensitic alloys [41,42].

(3) the D03 phase present in the temperature range
between 500 and 650°C can experience a spinodal
decomposition. The decomposition is believed to reach the
final stage after long aging, in which.the D03 structure
of the Ni3Sn or (Ni,Cu)3Sn phase would coexist with the
L21 structure of the CuzNiSn phase.

(4) the d2H phase is formed at temperatures below

500°C.

4.6 Differential Thermal Analysis

4.6.1 Results and Discussion

To delineate the phase diagram of (Ni,Cu)3Sn alloys,
a differential thermal analysis (DTA) method was used.
Specimens with Cu contents of 14, 17.5, and 20 at! were
examined by a DuPont Differential Thermal Analyzer after
rapid quenching from 1000°C. These specimens were heated
at a rate of 10°C/min in a dry argon atmosphere.

Figure 4.32a) shows a typical DTA trace obtained from
a specimen with a Cu content of 14 atZ Cu, whose starting
phase was an as-quenched 2H martensite. In this figure,
we observe four exothermic events during heating.
Essentially, similar events were observed for specimens
with Cu contents of 17,5 and 20 atZ, although their
starting phases were a retained D03 phase. DTA traces
obtained from the alloys having Cu contents of 17.5 and 20
atZ are shown in Figures 4.32(b and c). To see reaction
events more readily, transition temperatures seen in these
three figures are tabulated in Table 4.9 along with
reaction events that, we believe, took place during the
heating experiments.

The first exothermic reaction took place at 153°C in
the 14 atZ Cu alloy, and at 144°C in the 17.5 and 20 atZ

Cu alloys. This reaction has widely been accepted to

106

107

(e)
2141. 3"

75

 

120' 300 1.51 602 738 1 868
TEMPERATURE .—’ C’

Figure 4.32 AT—vs-T curves measured by a differential
thermal analyzer inNi-xCu-ZSSn alloys
during a heating process ( AT: temperature
difference). (a) x=14 atZ: the starting phase
being a quenched 2H martensite; (b) x=17.5
atZ: the starting phase being a D0 high
temperature phase; and (c) x-20 atg: the
starting phase being a D03 phase.

Table 4.9

Specimen
(starting phase)

Ni-l4Cu-258n
(ZH'martensite)

Ni-l7.5Cu-25$n
(retained D03)

Ni-20Cu-258n
(retained D03)

108

Summary of DTA analyses obtained from
as-quenched (Ni,Cu)33n specimens

Reaction
temperature (°C)

153
196
457
711
144

215
475

671
144

253
475

630

Reactions

migration of quenched-in
vacancies
2H'--->2H + d2H
d2H --->2H + D03
2H --->D03

migration of quenched-in

vacancies
D03 --—>2H + d2H

D0 -—->2H
d2 --—>2H + 003
23 —-->Do3

migration of quenched-in
vacancies
D03 --->2H + d2H
D0 —-->2H
d2 --->2H + D03
2H -—->D03

109

correspond to migration of quenched vacancies. Slight
differences in the reaction temperature indicate migration
energies different for various starting phases.

The second exothermic reaction startedtnzl96°C and
ended at 283°C in the 14 atZ Cu alloy. This reaction is
believed to correspond to the transformation from the
metastable 2H martensite phase (2H') to the stable 2H and
d2H phases. The second exothermic reaction observed in the
17.5 and 20 atZ Cu alloys occurred at 215 and at 253°C,
respectively. This reaction is considered to correspond
to the transformation from the retained D03 phase to the
2H and d2H phases.

The third exothermic reaction starting at
temperatures of about 457-475°C in the three alloys
corresponds to the transformation from the d2H phase to
the 2H and D03 phases. Between this temperature and the
temperature where the fourth reaction took place, we
believe that the dual phases, 2H and D03, were formed.
The fourth reaction occurred at 711°C in the 14 atZ Cu
alloy, at 671°C in the 17.5 atZ Cu alloy, and 630°C in the
20 atZ Cu alloy. This reaction corresponds to the
transformation from the 2H phase to the D03 phase. Above
this temperature and below melting point, all the alloys
used have the single phase of D03. By further heating the
alloys with Cu contents of 14 and 20 atZ, their melting

points were measured to be 1112 and 1022°C, respectively.

110

In the DTA experiments used here, we could not detect any
sign of spinodal decomposition reaction. This may be
caused bytacooling rate employed, which was faster than-

that needed for the decomposition.

4.6.2 Conclusions based on DTA Analysis

From the DTA analysis of the as-quenched alloy
specimens with Cu contents of 14, 17.5, and 20 atZ, the
four reactions were found to have taken place below their
melting points. The conclusions obtained in the present
heating experiments can be summarized as follows.

[1] In the metastable 2H martensite(2H'), vacancy
migration rate is highest at about 150°C, while the
corresponding maximum rate occurred at about 144°C in the
retained D03 phase.

[2] The 2H' martensite formed in the 14 atZ Cu alloy
was transformed at 200°C to the stable 2H phase and the
d2H phase.

[3] The retained D03 phase was transformed to the
stable 2H phase and the d2H phase at about 215°C in the
17.5 atZ Cu alloy and at about 250°C in the 20 atZ Cu
alloy.

[4] At temperatures of about 460°C, the
exothermic reaction from the d2H phase to the 2H and D03

phases took place in the three alloys with Cu contents of

111

14, 17.5, and 20 atZ.
[5] At temperatures of around 710’630°C, the 2H

phase was transformed to the D03 phase. The temperature
range between this temperature and melting point is

considered to be a region in which only the D03 phase is

present.

5 DISCUSSION

5.1 A Proposed Phase Diagram of the (Ni,Cu)3Sn System

The X-ray and electron diffraction analyses of the
alloy system have revealed that a deformed 2H phase was
formed both in furnace-cooled specimens of three Ni3Sn
alloys with Cu contents of 14, 17.5, and 20 atZ and in
specimens of the alloy with 20 atZ Cu aged at 350°C. In
the DTA analyses of as-quenched specimens of the three
alloys, it has been suggested that, in their heating-
experiments, the d2H phase was transformed to the D03 and
2H phases at about 460°C. In addition to these results,
the transmission electron microscopy'of‘alloy specimens
with a Cu content of 20 atZ has revealed that they
underwent a spinodal decomposition. The decomposition
probably occurs only in the temperature range between 460
and 700°C; the former temperature corresponds to the phase
boundary between a region of the D03 and 2H phases and one
of the D03, 2H, and d2H phases, according to the DTA
traces; the latter temperature corresponds to the phase
boundary between a single D03 region and the region of the
D03 and 2H phases.

In light of all the other results obtained in the
previous sections in conjunction with the above mentioned

results, we can propose a new phase diagram of the

112

113

(Ni,Cu)3Sn system, which is shown in Figure 5.1. In this
figure, the transition temperatures measured by the DTA
method were plotted as small solid circles. As seen in
this figure, the new phase diagram is quite different from
Figure 1.10 proposed by Murakami et al[7-9].

In Figure 5.1 a two-phase region of D03 and L21
phases was proposed to exist in the range of(h1contents
higher than 25 atZ, which was based on the following X-ray
data: Watanabe et a1 [7] reported that (Ni,Cu)3Sn alloys
had the D03 structure at temperatures below their melting
points for all alloys with Cu contents up to 25 atZ;
Klyucharev [39] found that a CuzNiSn phase had the L21
structure.

In the temperature range between about 700 and 460°C,

a region of three phases, D03, 2H, and L21, is proposed to

Footnote:

For the D0 structure, we can assume that Sn atoms
are located at a-sites and Ni and Cu atoms are located at
a-, c-, and d-sites at the same probability. Thus, the
structure factors for 111 and 002 ordered reflections can
be expressed as,

Fzoo ' F111 = fSn ‘ fNi ' 4(fCu ‘ fNi)xm/3

where xm is the molar fraction of Cu and f1 (i-Ni, Cu, Sn)
is an atomic scattering factor of the i-th atom.

0n the other hand, for the L21 structure, by assuming
that Sn atoms are located at a-sites, Ni.atoms.at c- and
d-sites, and Cu atoms at b-sites, the structure factors
can be expressed in the case where xIn is equal or less
than 0.25 as,

F111 ‘ fSn “ fNi + 4(fN1 ' fCu)xm/3

F200 ' fSn “ fNi ‘ 4(5N1 ‘ fCu)xm/3

114

 

 

 

 

 

 

 

l I I I l I " I
0‘0:.\.\_.~‘: : _
90.3 e
\ L2; L21
‘\\ DO3'2H E :d
\\ \ u I
\\____‘_______I___L—-\
/ I\ A
: \
2H*DO3*d2H : ‘. _
: ‘\
0 1 1 1 1 l J I]; 1
0 4 8 12 16 20 50
Ni-xCu-ZSSn x( at 'I.)
Figure 5.1

A proposed phase diagram for (Ni,Cu)3Sn alloys.

115

exist. Because the D03 and L21 structures are very similar
to each other, as seen in Figure 1.1, it is not easy to
differentiate them. The distinction can be made either by
measuring an intensity ratio of two given ordered
reflections, such as 111 and 200 (see footnote in the
previous page), or by chemical analysis through analytical
methods such as scanning transmission electron microscopy.
Although no identification of the L21 phase was attempted
in this present study, we could observe structural
evidence supporting the presence of the phase. The
presence of the L21 phase in this temperature range is

considered important in reasoning for the formation of
the d2H phase, which will be discussed in the next

section.

5.2 A Formation Mechanism of a Distorted 2H Phase

Electron microscopy and DTA analysis of the specimens
used in the present study have revealed that a D03 area
was transformed to both a 2H and a distorted 2H (i.e. d2H)
structure. The former is an ordered orthorhombic
structure (beta-Cu3Ti-type) and is t! stable phase of
(Ni,Cu)3Sn alloys with Cu contents ranging between 7 and
21 atZ. The same crystal structure is also seen in a
martensite (2H') transformed upon rapid quenching from the
D03 phase of (Ni,Cu)3Sn alloys with Cu contents up to 14
atZ. Why the orthorhombic 2H structure was formed from the
D03 phase in these two cases is explainable by taking atom
arrangements of their specific planes into account. 0n the
other hand, the latter is an‘ordered triclinic structure
and formation mechanisms for the lattice distortion of
the d2H structure are not clear if the D03 structure is
considered as a crystal structure of the parent phase.

In this section, therefore, we will discuss a
possible formation mechanism of the triclinic structure
originating from the D03 phase. This discussion will be
made after relationships of atomic arrangements are
described between the high-temperature D03 phase and its
three products, the stable and metastable 2H phases and

the D019 phase.

116

117

5.2fil Relationship Among D03, 2H, and D019 Structures

In Figure 5.2 are shown schematic atom arrangements
of close-packed planes for D019 and 2H structures of the
stoichiometric Ni3Sn alloy: (3) the (0001) basal plane of
D019 and (b) the (001) plane of 2H structure. By comparing
Figure 5.2(b) with Figure 1.1, one can notice that the
(001) plane of 8 2H martensite has an atom arrangement
similar to that of the (110) plane of the D03 parent
phase. This is because the transformation of the D03
phase to the 2H martensite does not take place by
diffusion of atoms but by shear motion of atoms. This
transformation is known to be explained by the Burger's
relation [43]. It should be noted from Figure 5.2(b) that
Sn atoms are surrounded by 6 Cu atoms holding two mirror
planes, perpendicular to each other: (100) and (010).
From this atom arrangement and because the third axis, c-
axis, is perpendicular to the (001) plane, it can be seen
that the 2H structure can hold an orthorhombic symmetry.
For a metastable 2H martensite formed in the alloy with a
Cu content of 14 atZ, an atom arrangement similar to that
shown in Figure 5.2(b) is expected; in this case, Sn
atoms are surrounded by a solid solution of Cu and Ni
atoms randomly occupying the 1“. sites in the
stoichiometric alloy, resulting in an orthorhombic

symmetry.

118

 

 

 

 

 

 

 

 

 

ea so '
@@@

Figure 5.2 Atomic arrangements of closed-packed planes.
(a) the (0001) plane of D019 structure;
(b) the (001) plane of 2H structure;
(c) the (001) plane of d2H structure.

119

For a stable 2H phase formed in both furnace-cooled
and aged alloys with Cu contents ranging between 7 and 20
atZ, a similar situation can be seen. In this case,
however, because of diffusion of atoms associated in
forming this 2H phase, the following situation is
possible: the compositions of Cu and Ni surrounding Sn
atoms are changed during the transformation from the D03
to the stable 2H phase but an: orthorhombic symmetry
remains unchanged by having a solid solution of Ni and Cu
around Sn atoms. This can be seen from the phase diagram
proposed in the previous section.

For the D019 structure, a similar discussion can be
made. This phase is massively transformed in alloys with
Cu contents up to 6 atZ. By comparing an atom arrangement
of the (0001) plane shown in Figure 5.2(a) with that in
Figure 1.1, one can find that the basal plane can be
formed by shear motion of atoms on every other (010)2H
planes by one atomic distance on the (001)2H plane. Note
that Sn atoms are surrounded by 6 Ni atoms holding a six-
fold symmetrical axis perpendicular to the basal plane.
That is, the D019 structure has a hexagonal structure but
can be considered to be an orthorhombic structure. The a-
and b-axes, when this orthorhombic lattice is considered,
are shown in Figure 5.2(a). Comparing axial ratios of b/a
for the 2H and D019 structures, one can notice that the 2H

structure is slightly distorted from the hcp symmetry.

120

5.2.2 Formation of d2H from D03 via L21

The above mentioned examples show that the
orthorhombic symmetry of the 2H and D019 structures can
be derived from the atom arrangements of the (110) plane
of the D03 structure. This suggests that the parent phase
of the d2H should not be the D03 phase. Further
distortion from the 2H lattice, when the d2H lattice is
formed, can be explained based on the assumption that some
compositional changes occur in the vicihity of D03 regions
wherein the stable 2H phase is formed.

As discussed in the previous section, an L21 phase
can be transformed from the high-temperature D03 phase by
the following processes: 1) a spinodal decomposition of
the D03 phase, and 2) a diffusion process associated with
the formation of an Ni-rich 2H phase. The former process
would not occur in the case when specimens are furnace-
cooled, while the latter process seems more plausible.

Figure 5S3 shows schematic illustrations describing
the formation process of a region of L21 structure. As
shown in the phase diagram (Figure 5.1), a specimen of
the D03 structure is stable at temperatures higher than
about 760°C. As soon as the temperature of the specimen
is lowered below 760°C, stable 2H regions start to form
from the D03 phase. 'These 2H regions should be Ni-rich,

and consequently, areas surrounding them become Ni-poor

121

 

 

 

 

 

 

 

 

 

 

High
0'.
E ---'L2]
m
'—
v 1’ Ni-rich 2H formed
Low from 003

 

  

 

 

 

[Cu-rich deformed 2H
formed from L2]

Figure 5.3 Schematic illustration describing a formation
process of a d2H from D03 via L21.

122

and Cu-rich. The resultant Cu-rich areas can change or
can be transformed to the L21 phase, because this
structure bears great resemblance to the D03 structure.
When the temperature is further lowered below 470°C, 8 new
phase is formed in the L21 region and can be distorted due
to an atom arrangement of the (001) plane of the new (i.e.
d2H) phase because of the following reason.

Figure 5.2(c) shows such an atom arrangement of the
(001) plane of the d2H structure. Comparing Figures
5.2(a,b, and c), one can notice that the crystal symmetry
of the basal plane becomes the highest for the D019
structure and the lowest for the d2H structure. Although
the two-fold symmetry of the (001)2H plane still remains
unaltered in the (001) plane of the d2H phase, a further
distortion from the 2H lattice is considered to take place
in Figure 5.2(c). It is well known that some alloys such
as VIr and MnAu[44] have a tetragonality from CsCl-type
structure even though they have high lattice symmetry as
long as the rigid atom model is concerned(Table 5.LL
Therefore, it is likely that the crystal structure formed
from-thelLZI phase can undergo further lattice distortion,

resulting in the triclinic structure of the d2H phase.

123

Table 5.1 Examples of tetragonally distorted CsCl—type

compounds.
VRU c/a= 1.04
VIr c/a= 1.34
TiRh c/a= 1.14
Tilr c/a= 1.17
MnAu c/a: 0.95

ZnNi c/a= 1.16

 

6 SUMMARY AND CONCLUSIONS

The phase transformations:ha(Ni,Cu)3Sn alloys were
investigated in detail by optical and scanning and trans-
mission electron microscopy, electron and X-ray diffrac-
tion techniques, and differential thermal analysis. The
alloys used were the ones containing Cu contents ranging
between 10 and 22 atZ. The results obtained in this study
can be summarized as follows:

[1] Specimens of alloys with Cu contents of 14,
17.5, and 20 atz were furnace-cooled from 1000°C and it
was found that those specimens showed the presence of two
phases. One was the most abundant phase, exhibiting rod-
shaped microstructures, and the other was the less
abundant phase, exhibiting very fine acicular structures.
Electron and X-ray diffraction analyses and electron
microscopy manifested that the former phase had an ordered
2H (beta-Cu3Ti-type) orthorhombic structure, containing
(101) faults and a low density of dislocations. The
latter phase was identified to have a triclinic structure,
slightly distorted from the 2H structure, denoted as d2H.
The acicular structures of this d2H phase contained a
large number of uniformly distributed internal planar
faults. The lattice parameters measured of the d2H phase
were a-4.53A, b-5.31A, c=4.34A, alpha-85°, beta-86°, and
gamma-84°. X-ray diffraction analyses showed that the

three phases of 2H, d2H, and D03 were present in furnace-

124

125

cooled specimens.

[2] Several (Ni,Cu)3Sn alloys were quenched from
1000°C to ice water. It was found that a high-temperature
D03 phase was retained for as-quenched specimens contain-
ing Cu contents ranging from 16 and 22 atZ. This observa-
tion disagrees with results obtained by Murakami et a1
[7-9] who reported that the high-temperature phase was
transformed to a 2H martensite» In this study, such a 2H
martensite was only observed in specimens with a Cu
content of 14 atZ. This martensite was found to contain
(121) twins.

[3] Aging experiments of a Ni3Sn alloy containing a
Cu content of 20 atZ were carried out for 24 hr at temper-
atures of 350, 500, 600, and 650°C. From the four aging
experiments, the following conclusions can be made.

(a) The high temperature D03 phase was stable above
650°C and below the melting point (1022°C) of this alloy;
however, this phase was transformed in part to the 2H
phase between 500 and 650°C to be the dual phases, the D03
and 2H phases.

(b) The D03 phase appears to be metastable at
temperatures of about 650°C. From the appearance of the
X-ray line-broadening, the D03 phase is considered to have
a pre-transformation, as is often observed in many ordered
martensitic alloys.

(c) The D03 phase present in the temperature range

126

between 500 and 650°C can experience a spinodal
decomposition. The decomposition is believed to reach the
final stage after long aging, in which the D03 structure
of the N138n or (Ni,Cu)3Sn phase would coexist with the
L21 structure of the CuzNiSn phase.

(d) The d2H phase should be formed at temperatures
below 500°C.

[4] From differential thermal analyses and electron
diffraction analyses, the following results were obtained.
The high-temperature D03 phase was transformed to an
ordered 2H (beta-Cu3Ti-type) orthorhombic phase at
temperatures ranging between 710 and 630°C in alloys
containing Cu contents between 14 and 20 atZ. Residual
areas of the D03 phase were then transformed to both the
2H phase and the d2H phase at a temperature of around
460°C.

[5] Based on the experimental results obtained in
this study, a new phase diagram was proposed for the
(Ni,Cu)3Sn alloys in the range of Cu contents between 10
and 22 atZ. This phase diagram was different from the one
available in literature, which shows a single phase of 2H
at temperatures below about 900°C in the range of Cu
contents between 8 and 18 atZ.

[6] It is proposed that the d2H phase is formed at
about 460°C from an L21 phase which is considered to be

transformed in residual D03 regions neighboring 2H areas.

REFERENCES

1)
2)

3)

4)

5)
6)

7)

8)
9)
10)

11)
12)

13)
14)
15)
16)

17)

18)
19)

20)

S. Rosen, J. A. Geobel: Trans. AIME, 242(1968), 722.

K. Enami,S. Nenno, andK. Shimizu: Trans. JIM,
14(1972), 161.

J. L. Smialek and R. H. Heheman: Met.Trans,
4(1973), 1571.

S. Chakravorty and C. M. Wayman: Met. Trans.,
7A(l976), 569.

Y. Murakami and S. Kachi: Trans. JIM, 18(1977), 423.

H.-r. Pak, T. Saburi, and S. Nenno: Trans. JIM,
37(1973), 1128.

Y. Yatanabe, Y. Murakami, and S.Kachi: Trans. JIM,
6(1981), 551.

Y. Murakami and S. Kachi: Trans. JIM, 24(1983),9.
Y. Murakami and S. Kachi: Trans. JIM, 24(1983), 747.

K. Otsuka, T. Sawamura, and K. Shimizu: Phys. Stat.
$01, (a) 5(1971), 457.

K. Otsuka and K. Shimizu: Phil. Mag., 24(1971), 481.

D. B. Masson andC. S. Barrett: Trans. AIME,
212(1958),260.

D. B. Masson: Trans.AIME., 218(1960), 64.
L. C. Chang: Acta Cryst., 4(1951), 320.
R. S. Toth and H. Sato: Acta Meta11., 16(1968), 413.

P. L. Ferraglio and K. Mukherjee: Acta Metall..
22(1974), 835.

T. B. Massalski and H. W. King: Progress in
Materials Science, 10(1961), 1.

H. Sato: Met. Soc. Conf., 29(1963), 295.

H. Sato, R. S. Toth, and G. Honjo: Phys. Chem. Solid,
28(1967), 137.

S. Kajiwara: J. Phys. Soc. Japan, 22(1967), 795.

127

21)
22)
23)

24)

25)

26)

27)

28)

29)

30)

31)

32)

33)

34)
35)
36)

37)
38)

39)

40)

128

A. B. Greninger: Trans. AIME., 133(1939), 204.
T. B. Massalski: Acta Metall., 6(1958), 243.

J. E. Kittl and T. B. Massalski: Acta Metall.,
15(1967), 161.

K. H. G. Ashbee, L. F. Vassamillet, and T. B.
Massalski: Acta Met., 15(1967), 181.

T. Saburi and C. M. Wayman: Trans. AIME, 233(1965),
1373.

J. F. Kittl and C. Rodriguerz: Acta Metall.,
17(1969), 925.

M. Hansen and K. Anderko: Constitution of Binary
Alloys, McGraw-Hill, New York, (1958), 1042.

O. Nial: Svensk Kem. Tidskr., 59(1947), 172.

KL Schubert,1% Esslinger, E. Gunzel, G. Meissner,
W. Schtt, J. Wegst, and M. Wilkens: Naturwiss,
43(1956), 248.

S. N. Chang: Matser Thesis, MSU, 1983.

H.-r. Pak, S. N. Chang, and M. Kato: J. Mat. Sci.,
20(1985), 947.

K.Osuka and K. Shimizu: Jap. J.Appl. Phys.,
8(1969), 1196.

H. Kubo, I. Cornels, and C. M. Wayman; Acta. Metall.,
28 (1980), 405.

M. Bouchard and G. Thomas: Acta Metall., 23(1975), 1485.
P. Rahlfs: Metallwirtschaft, 16(1937), 343.

H. Nowotny and K. Schubert: Z. Metallkde,
37(1946), 17.

L. Castelling: Mh. Chem., 84(1953), 765.

L. A. Carapella and R. Hultgren:Phys. Rev..
59(1941), 905.

A. P.K1yucharev: Z. Eksper. Teort. Fiz..
9(1939), 1501.

H. Hendus and H. Knodler: Acta Cryst., 9(1956), 1036.

41)

42)

43)
44)

129

L. Delaey, A.J. Perkins and T.B. Massalski:

J. Mat. Sci., 7(1972), 1197.

C. M. Wayman: Mat. Res. Soc. Symp. Proc.,

39(1985), 77.

W.(L Burgers: Physica,1(1934),561.

A.ILDwight: Intermetallic Compounds,

Westbrook, John Wiley & Sons,

Inc.,

NY,

ed. J.H.
1967.