rapid solidification processing of indium gallium antimonide alloyssecure site · 2020. 4. 2. ·...
TRANSCRIPT
RAPID SOLIDIFICATION PROCESSING OFINDIUM GALLIUM ANTIMONIDE ALLOYS
Item Type text; Thesis-Reproduction (electronic)
Authors Kumta, Prashant Nagesh, 1960-
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this materialis made possible by the University Libraries, University of Arizona.Further transmission, reproduction or presentation (such aspublic display or performance) of protected items is prohibitedexcept with permission of the author.
Download date 17/04/2021 03:54:05
Link to Item http://hdl.handle.net/10150/276468
INFORMATION TO USERS
This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality of the material submitted.
The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.
1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.
2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of "sectioning" the material has been followed. It is customary to begin filming at the upper left hand corner of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.
4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.
University Microfilms
International 300 N. Zeeb Road Ann Arbor, Ml 48106
Order Number 1331410
Rapid solidification processing of indium gallium antimonide alloys
Kumta, Prashant Nagesh, M.S.
The University of Arizona, 1987
U M I 300 N. ZeebRd. Ann Arbor, MI 48106
PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark V .
1. Glossy photographs or pages ^
2. Colored illustrations, paper or print
3. Photographs with dark background
4. Illustrations are poor copy
5. Pages with black marks, not original copy s/
6. Print shows through as there is text on both sides of page
7. Indistinct, broken or small print on several pages
8. Print exceeds margin requirements
9. Tightly bound copy with print lost in spine
10. Computer printout pages with indistinct print
11. Page(s) lacking when material received, and not available from school or author.
12. Page(s) seem to be missing in numbering only as text follows.
13. Two pages numbered . Text follows.
14. Curling and wrinkled pages
15. Dissertation contains pages with print at a slant, filmed as received
16. Other
University Microfilms
International
RAPID SOLIDIFICATION PROCESSING OF
INDIUM GALLIUM ANTIMONIDE ALLOYS
by
Prashant Kumta
A Thesis Submitted to the Faculty of the
DEPARTMENT OF MATERIALS SCIENCE AND ENGINEERING
In Partial Fulfillment of the Requirements For the Degree of
MASTER OF SCIENCE
In the Graduate College
THE UNIVERSITY OF ARIZONA
19 8 7
STATEMENT BY AUTHOR
This thesis has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the Uhiversity Library to be made available to borrowers under rules of the Library.
Brief quotations from this thesis are allowable without special permission, provided that accurate acknowledgement of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department of the Dean of the Graduate College when in his or her judgement the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.
SIGNED:
APPROVAL BY THESIS DIRECTOR
Tl is has been approved on the date shown below:
- h • 7- /£> — 8 ~7 S. H. Risbud
Professor of Materials Science and Engineering
Date
To my dearest Parents "Nagesh and Sumati Kumta"
ACKNOWLEDGEMENTS
I wish to thank my advisor, Dr. Subhash Rlsbud for the constant
encouragement, advice, and assistance needed to complete this study. I
also wish to express my gratitude to Dr. David Poirier for his valuable
suggestions during the writing of this thesis. Sincere thanks are also
extended to Mr. Dick Van Reeth for his help in machining the melt
spinning set up needed for this study. I am also grateful to my dear
friend Mr. V. P. Dravid who performed electron microscopy on the
rapidly solidified flakes with the microscopy facility at Lehigh
University. The help of Mr. D. S. Mishra of the Pharmaceutical Science
Department at the University of Arizona who performed the DSC analysis
is also appreciated. Finally, I would like to thank my parents, who are
mainly responsible for what I have accomplished.
ill
TABLE OF CONTENTS
Page
LIST OF ILLUSTRATIONS v
LIST OF TABLES x
ABSTRACT xi
CHAPTER
1. INTRODUCTION 1
2. BACKGROUND AND LITERATURE 4
2.1 Rapid Solidification Technology 4 2.2 III-V Semiconductors 21 2.3 Ternary III-V Antimonides 21 2.4 III-V Ternary phase diagrams 26 2.5 Synthesis Of Bulk Crystals Of
InxGa1_xSb 28
2.6 Thin Film Methods 33
3. EXPERIMENTAL SET UP AND PROCEDURE 35
3.1 Materials Preparation 35 3.2 Materials Characterization 40
4. RESULTS AND DISCUSSION 44
4.1 X-ray Diffraction Analysis 44 4.2 Surface Morphology 53 4.3 Thickness Measurements 63 4.4 Microstructure Analysis 67 4.5 Transformations 85 4.6 Electrical Property Measurements 93
5. SUMMARY AND CONCLUSIONS 97
REFERENCES 99
iv
LIST OF ILLUSTRATIONS
Figure Page
1. Cooling rates and various structural forms achieved in rapid quenching from the liquid state 5
2. Schematic for representing the concept associated with undercooling of the a phase for a simple eutectic system 7
3. Extended phase concept for adiabatic solidification of the a phase 9
4. Schematic concept for undercooling development of metastabler phase and a special case of hyper cooling where 7 phase was suppressed leading to formation of an amorphous phase 11
5. Microstructural consequences of rapid solidification 12
6. Schematic representation of various techniques of RSP 17
7. Various techniques of filament casting 19
8. Powder preparation techniques 22
9. The InSb-GaSb pseudobinary phase diagram 29
10. Schematic diagram of synthesis of InxGa1_xSb alloys 36
11. Schematic of the RSP set up 38
12. X-ray trace of ingot, flake and powder obtained from melt of run 19 50
v
vi
LIST OF ILLUSTRATIONS — Continued
Figure Page
13. (a) SEM micrograph showing surface of flake obtained from run 9 54
(b) SEM micrograph showing surface of flake obtained fran run 20 54
(c) SEM micrograph showing precipitates in flakes obtained form run 1 55
14. ED AX analysis of the precipitate observed on the flake surface 57
15. EDAX analysis of the matrix on the flake surface . 58
16. (a) SEM micrograph of powder obtained from run 10 59
(b) SEM micrograph of powder obtained from run 10 59
(c) SEM micrograph of powder obtained from run 10 60
(d) SEM micrograph of powder obtained frcxn run 10 60
(e) SEM micrograph shewing fractured powder particle obtained fran run 10 61
(f) SEM micrograph shewing powder obtained from run 9 61
(g) SEM micrograph showing powder obtained from run 8 62
17. (a) SEM micrograph of cross section of flake obtained from run 1 64
(b) SEM micrograph of cross section of flake obtained from run 19 64
18. Variation of flake thickness with disk velocity 65
19. Plot of experimentally determined flake thickness (d, microns) versus cooling rate ( K/s) calculated assuming validity of the equation Cd =81 68
4
vii
LIST OF ILLUSTRATIONS — Continued
Figure Page
20. (a) Optical micrograph showing cross section of ingot of composition IriQ 46GaQ 54Sb 70
(b) Optical micrograph showing the etched surface of the flake obtained fran run 7 70
(c) SEM micrograph shewing the etched surface of the flake observed in Figure 20(b) 71
21. (a) Optical micrograph showing the cross section of flake obtained fran run 7 73
(b) Optical micrograph showing the cross section of flake obtained from run 14 73
(c) SEM micrograph showing the cross section of flake observed in Figure 21(b) 74
22. (a) SEM micrograph showing the etched cross section of powder obtained from run 14 75
(b) Optical micrograph showing the etched cross section of powder obtained from run 14 75
(c) Optical micrograph showing the etched cross section of pcwder obtained from run 10 76
23. (a) Optical micrograph showing the etched surface of the flake obtained frcm run 18 78
(b) Optical mocrograph shewing etched surface of the flake obtained from run 20 78
(c) optical micrograph showing the etched surface of the flake obtained from run 19 79
24. SEM micrograph of etched surface of flake observed in Figure 23(a) 79
25
26
27
28
29
30
31
32
33
34
35
36
LIST OF ILLUSTRATIONS — Continued
Variation of grain size with disk velocity
Bright field image of flake obtained from run 13
Diffraction pattern obtained frcm an area of flake observed in Figure 26
Diffraction pattern obtained for powder obtained from run 20
Bright field image of the same region from which diffraction pattern in Figure 28 was obtained
Diffraction pattern obtained for powder obtained from run 20 after exposing the sample to the electron beam for 10 minutes .
Dark field image of the region from which diffraction pattern in Figure 28 was obtained
Auger electron spectra showing presence of C, 0, Cu, In, Ga, Sb in flake obtained frcm run 16
Depth profile for Cu, C and elemental 0 in flake obtained from run 16
DSC traces of ingot and flakes obtained from run 18, 19 and 20
DSC traces of corresponding powders obtained from run 18, 19 and 20
DSC traces of powder obtained from run 8, 9 and 10
viii
Page
80
82
82
83
83
84
84
86
87
88
90
92
1
ix
LIST OF ILLUSTRATIONS — Continued
Figure Page
37. Variation of bulk electrical resistivity with annealing temperatures for flakes obtained from run 7 94
LIST OF TABLES
Table Page
1. Range of dendritic spacings associated with cooling rate and processes 14
2. Sunmary of cooling rate measurements for splat cooling 15
3. Quench rates and sample shapes in splat quenching 18
4. Fiber processing techniques 20
5. Powder processing techniques 23
6. Selected properties of III-V compounds, germanium and silicon 24
7. Possible combinations of the III-V antimonides 25
8. Lattice parameters of binary constituents in III-V ternary antimonides 27
9. Melt spinning compositions with melt spinning parameters 45
10. Lattice parameter values 49
11. Summary of XRD and DSC analysis 91
12. Resistivity values cited in literature 96
x
ABSTRACT
Solidification from the melt is an essential step in nearly all
conventional processes to produce bulk materials for industrial
applications. Rapid quenching from the liquid state at cooling rates of
102 to 106K/S or higher has developed into a new technology for
processing novel materials. I^Ga^.^b a ternary III-V compound
semiconductor was synthesized by using the rapid spinning cup (RSC)
technique. Several compositions of these alloys were batched and cast
into ingots in evacuated sealed quartz tubes. These ingots were then
melted and ejected onto a rapidly rotating copper disk. This resulted
in the generation of flakes or powders depending on the rpm of the
disk. Mlcrostructural characterizatlcn of the flakes and powders was
performed using XRD, SEM and TEM. Efforts were also made to measure the
bulk resistivity of the annealed flakes to see the effect of annealing
on ordering of the phases.
xi
CHAPTER 1
INTRODUCTION
The Increasing Interest In the field of microelectronics has
led to a vigorous and rapid development of semiconductor materials
science, technology and device fabrication. The last several years have
witnessed the large scale application of silicon in new devices with
enormous size reductions and packing densities. The real emergence of
silicon as a semiconductor material Is based cn Its unique advantages,
although It possesses certain limitations. The band gap of silicon and
germanium are indirect and are fixed at definite values of Eg(Ge) =
0.68 eV and Eg(Si) * 1.1 eV. Owing to these factors they are not useful
as semiconductor light sources or microwave devices; the energy band
structure Is suitable neither for light emitting diodes and lasers, nor
for transferred electron devices ( Gunn diodes ). This is the reason
why the interest in compound semiconductors is growing continuously
with an examination of their capability for electronic applications[ 1 ].
Among the possible compound semiconductors the III-V compounds
are the most Important ones. The group consists of compounds which are t
composed of the group IIIA and the group VA elements, however the
compounds containing boron and nitrogen are generally omitted primarily
due to their physical and chemical nature. The growth of large single
1
2
crystals, epitaxial layers or p-n junctions of these materials is very
difficult. Combination of these six remaining elements ( P,As and Sb in
the group V and Al, Ga and In in the group III ) gives nine binaries,
eighteen ternaries and fifteen quaternary compounds. Only parts of
these combinations are known yet, and among them one can find compounds
as GaAs, GaSb, InSb, GaP, GaAsP, InGaAs, GaAlAs or GalnAsP which have
great technical importance forming the material base of industrial mass
production (LED's, GaAs devices etc ) and optoelectronic and microwave
devices.
The III-V compounds are used in optoelectronics and microwave
technologies. Furthermore new applications of these compounds such as
high speed digital integrated circuits (switching rates of less than
lOOps at very low power level) are also promising. At present a large
amount of data for III-V binaries are available but interest has
increased recently in some mixed III-V alloy systems. Compared to the
continuously growing effort in developing appropriate technological
methods, for III-V binary compounds, the preparation techniques for
ternary and quaternary III-V compounds have yet to fully exploit the
emerging rapid solidification technology currently used mainly for
metallic systems.
The present work is thus focussed on rapid solidification
processing (RSP) of Ir Ga ^ Sb alloys, of interest in galvanomagnetic
devices and as infra red detectors in the spectral range 1.8 - 6.9 fim.
3
The major objectives of the research were synthesis and
characterization of rapidly solidified flakes and powders using X-ray
diffraction (XRD), scanning electron microscopy (SEM), transmission
electron microscopy (TEM) and Auger electron spectroscopy (AES)
techniques.
CHAPTER 2
BACKGROUND AND LITERATURE
2.1 Solidification Technology
The technique of quenching or rapid cooling in the solid state
has been practiced fay metallurgists and material scientists [2-5] in
order to develop suitable microstructures and obtain desired
properties. However quenching from the liquid state in order to modify
the process of solidification to obtain a solid has in the recent years
developed into a new technology known as Rapid Solidification
Technology (RST).
The cooling rates and the corresponding various structural
forms achieved In rapid solidification is Indicated in Fig.l. Starting
from the liquid state [6], a cooling rate of 106K/s or higher can yield
metallic glasses. Cooling rates in the range from 102 to 106K/s produce
crystalline alloys, which are called, depending an their average size,
micro or nano crystalline. Hie rapidly solidified materials have the
form of thin ribbons, wires, flakes or powders.
It is currently recognized that rapid solidification affords i
numerous opportunities in materials development.[7]. The primary
virtues afforded by rapid solidification are a) Chemical homogeneity,
or decreased chemical segregation, b) refined grains and other
4
LIQUID STATE |
METALLIC 10* GLASSES
.10* NANO MICRO
QUENCHING CONVENTIONAL METALLURGY
.9* -
COOUNG RATE (Ks" I
Figure 1. Cooling rates and various structural fox'ms achieved in rapid quenching from the liquid state.[6]
6
structures, c) increased solute solubility above conventional
equilibrium levels and d) metastable phase development which includes
amorphous and microcrystalline structures.
The concept of chemical segregation development during rapid
solidification of an alloy accompanied by extensive undercooling can be
visualized from a simple eutectic phase diagram shown in Fig.2. This
model illustrates the concept of undercooling, where nucleation of the
solid phase sets in at temperatures significantly below the liquidus
temperature (Tj) of the alloy at hand. Two assumptions are applied to
the model a) heterogeneous nucleation sites are minimal and b)
adiabatic conditions, i.e the latent heat of solidification AH is
absorbed fast enough within the system so that, effectively, any heat
removal from the system during solidification is inconsequential. The
model shown in Fig.2 illustrates three undercooling conditions from the
liquid state and is associated with temperatures T1( T2 and T3. The
degree of undercooling is associated with the recalescence temperature
Tr (temperature obtained from the release of latent heat of
solidification), and the solidus temperature Tg.
Case l.[7]: hypocooling where undercooling from the liquid state
to Ta occurs but the latent heat of solidification raises the
temperature above the solidus temperature, i.e TR > Tg. In this
instance, local equilibrium and solute partitioning occurs since Tr is
in the a + L field. This condition leads to dendritic growth with
solute rich cores, i.e. chemical segregation. The dendrites become
Of + B
V.B
Increasing undercooling
Critical Hypocooling undercooling Hypercooling
Homogeneous » Segregated Homogeneous a
Time Time Time
Figure 2. Schematic for representing the concept associated with undercooling of the a-phase for a simple eutectic system. [7]
8
larger as TR increases.
Case 2T71: Undercooling from the liquid state to T2 and in
this instance TR = Tg. This condition is described as critical
undercooling and is associated with
AH =T2/TsCp1 dT (1)
where A H is the latent heat of solidification, and Cp1 is heat
capacity of the liquid phase. In this case solidification can occur
massively within the a phase without compositional piartitioning and
local equilibrium at the solid/liquid interface.
Case 3[7]: hypercooling and is represented by TR < Tg and,
similar to the previous condition, results in complete solute trapping
hence segregationless. solidification. As a result the composition of
the a phase is uniform and is equal to CQ. The last two conditions
feature very rapid nucleation, and as a result fine or microcrystalline
grains are expected to dominate the morphology.
The same concept of undercooling and recalescence can be
extended to the understanding of the formation of metastable phases.
This concept induced by rapid solidification is shown in Fig. 3 where
once again a simple binary eutectic is used for the illustration.
Again, adiabatic solidification is assumed. If the melt can be
sufficiently undercooled, the liquidus and solidus boundaries can be
extended such that the two phase region of « + L in Fig.3 will coexist
at temperatures below the equilibrium eutectic leading to a condition
T1
%B »-
Figure 3 . Extended phase concept for solidification of thea-phase.[7]
adiabatic
<0
10
where the /3 phase is totally suppressed. In principle conditions of
hypocooling, critical undercooling and hypercooling can be generated
(see Fig.2) producing a range of extended a phase morphologies i.e.
dendritic (segregation), cellular, and completely homogeneous. The
degree to which phases, such as a shown .in Fig.3 can be extended using
current rapid solidification techniques, is probably not large but
significant.[7,8]
Other forms of metastable phase development afforded by rapid
solidification can lead to new and unique phases. An example of such
behaviour is shown in Fig. 4, where the metastable r phase forms at the
expense of the more stable a and 0 phases as a result of appropriate
conditions of undercooling. If a condition of critical undercooling
occurs, the metastable y phase can form in a homogeneous manner. Also,
it is conceivable that if the amount of undercooling, i.e. below the
glass transition temperature Tg, is very large then the heat of
solidification, AH, is zero, and the metastable r phase can be
suppressed. This would lead to the formulation of an amorphous phase.
Microstructural consequences of rapid cooling from the molten
state are summarized in Fig.5. In passing from ordinary casting
practice to cooling rates greater than 102 K/s, the microstructural
features become refined because the time for coarsening during
solidification is reduced. [8]
The morphological assessment of rapidly solidified materials
has been mainly associated with dendritic behaviour. Jones[9],
Increasing undercooling
Figure 4. Schematic concept for undercooling development of metastable 7 phase and a special case of hyper coo ling where 7 phase was suppressed leading to formation of an amorphous phase.[7]
rr Coarse dandritas, eutectlcs and other mlcroconstltuants
Incraasing cooling rata (Kfs)
Refined microstructures
Composition and process Fine dendrites,
eutectlcs and other mlcroconstltuents
Refined microstructures dependent r
Fine dendrites, eutectlcs and other mlcroconstltuents
4J-Novel microstructuraa
Extended solid solutions, mlcrocrystalllne structures, metastable crystalline phases, amorphoua solids
Novel microstructuraa
Extended solid solutions, mlcrocrystalllne structures, metastable crystalline phases, amorphoua solids
Increasing homoganalty
Figure 5. Mlcrostructural consequences of rapid solidification.[8]
13
Mehrabian[10], Grantll] and Flinn[7] have provided a correlation for
dendritic spacings provided by solidification from the melt as a
function of cooling rate. This correlation is shown in Table 1. A
detailed discussion on the determination of the cooling rate from the
microstructures obtained by rapid solidification can be found in the
book written by Flinn[7]. However, it has also been possible to
estimate the cooling rate from the specimen thickness. Jones[12] has
discussed in detail the various types of cooling conditions that can be
obtained and the corresponding agreement of the theoretical and
experimental values of the cooling rate. The available experimental
measurements of cooling rate by Predecki et al.[13], Harbur et al.
[14], Miroschnichenko and Zakharov [15], Lohberg and Muller [16,17],
Brower et al. [18] are summarized in Table 2. They range from about 105
- 10® K/s for splat thicknesses ranging from about 100/zm down to 1 jum.
Ruhl [19] has also pointed out that the estimation of cooling rates
depended on the cooling conditions being Newtonian, intermediate, non-
Newtonian and ideal. He defined the condition of ideal cooling as the
state of ideal or perfect thermal contact between splat and substrate
giving rise to the condition of an infinite heat transfer coefficient
between the splat and substrate.
Under these conditions he reported that the cooling rate (T «
1/x ) where ~x' is the splat thickness. His calculations showed that
variations of splat temperature, solidification temperature and
substrate temperature had little effect on T compared with likely
14
Table 1. Range of dendritic spacings associated with cooling rate and processes [7].
Cooling Ratt Characteristic Characteristic
Uniting Oendrltlc K/s Designation Processes Thickness* Spacing
£ 3 10 to 10 Very low Large sand cast- >6 m 0.S to 5 m
Ings and Ingots
10"* to 10° Low Standard castings 0.2 to 6 m SO to 500 vm and Ingots
lo" to 10^ Medium-High Thin strip, die 6 to 20 nm 5 to 50 ym casting, conventional atomlzatlon
103 to 106 High Fine powder atom- 0.2 to 6 nn 0.5 to S ym Izatlon, me It extrusion/extraction
10* to 103 Ultra-High Spray deposition, 6 to 200 va 0.0S to 0.S m and higher melt spinning,
electron/laser beam glazing
a. Assuno geometry is a metal slab from which the heat Is readily extracted from principal surfaces.
Table 2. Summary of cooling rate measurements for splat cooling [12]
Location of measurement
Interface with substrate
Within splat
Exposed splat surface
Splat technique
Gun
Piston and anvil
Piston and anvil
Two-piston
Gun
{ {
Splat
Ag Al Au-14"/0Sb Fe alloy Al l'b Pb-50w/o Sn Al-4w/0 Cr
Al Cii and Cu-Ni
} } }
Substrate
Ni/Afj
Fused silica
Fe/Al
Mica on Cu
Cu
{
Thickness of splat s (urn)
~ 1
~27J 701 44t 44J 65t
~20
{
Initial cooling rate 'i'(Ks ')t
1 xl0"-5xl0« 1-5 x 10'-3 x 10' ~10s
up to 10" 6-9 x 10* 1-2 x 10* 7-8 x 10s
9 7 x 10'
^ up to 1-5 x 10*
} }
up to 10s
Reference
Prcdecki et al
Hrower ct al [18]
llarbur el al [14]
Miroshnichcnko and [15] Zakharov
Liililierj; and Miillcrf 15^
t Cooling rate '1' was measured tlicnnoclcctrically cxccpt by Hrower ct al and Lfihberg and Mi'tllcr who used optical pyrometry. llarbur el al and Lfihberg and Miiller were able to splat cool under an inert atmosphere.
} Half thickness (twin substrates).
16
variations in ~ h1 and' s1 where 'h' is the heat transfer coefficient
between the splat and the substrate and ~s' is the splat thickness.
References[12,20-24] also provide the reader with an excellent
comparison of the cooling rates predicted by theory and the values
determined experimentally.
There are several techniques of RSP described in the
literature. Fig.6 shows a representative set of liquid quench [25]
procedures all closely related, which are commonly used in laboratory
preparation of metastable alloys. A description of the various
processes and their operation can be found in referencesf22,25-33].
Each of these processes result in the formation of flakes, foils or
tapes. Table 3 lists these processes and compares the cooling rates and
the final products obtained in each of the process. The methods
illustrated in Fig.6 are generally used to produce short experimental
specimens in either rod, strip or flake form. However, some of them,
for example the processes shown in Fig.6(d), 6(e) and 6(f) can however
be adapted to produce uniform continuous lengths of rapidly chilled
fibre or ribbon, or a continuous stream of staple fibre or powder like
material. Melt spinning and melt extraction, the relative quench rates
of which have been discussed by Pond and Winter [34] are two such
modifications. There are several other fiber processing techniques
illustrated in Fig.7. These methods are classified according to the
ways in which liquid metal is supplied to, and heat removed, from, the
solidification surface[35-39]. Table 4 lists these processes in the
Figure 6. Schematic representation of various techniques of RSP.[25]
Table 3. Quench rates and sample shapes in splat quenching [25].
Me thod
Cooling rate (#C sec"*) Sample shape
I Gun
Piston and anvil
Plasma-jet
(Torsion catapault
(Drum quenching
Roller quenching
106 to 108
106 to 107
~ 10
~ 106
~ 10€
~ 10
Irregular thin film or powder (rather uniform thickness)
Irregular thin plates (uniform thickness)
Irregular thin film (uniform thickness)
Irregular thin plates (uniform thickness)~
Regular uniform ribbon filaments
(uniform thickness)
Regular thin plates (uniform thickness)
19
FREttUW
(a)
(c)
(e)
metal roo
QUENCHING MEDIUM AREA
souo wme
WATER-FILAMENT
(b)
(d)
( f )
Figure 7. Various techniques of filament casting: (a) melt drawing sheathed In glass; (b) melt extrusion Into a stabilizing medium; (c) melt spinning from a rotating chill surface;(d) melt-rolling (twin-roll quenching); (e) melt extraction from a crucible; and (f) melt-drag from a crucible side nozzle.[35]
Table 4. Fiber Processing Techniques
Process Cooling Rate Productof solidification
1.Taylor's Wire 104-106K/s wires 2-50/im dia. Process[36]
2. Melt 105-108K/s filaments 1-100/im. Spinning [34] thick.
6 T 3.Rotating 10 -10 K/s ribbons 3-Bim. Cup[36]
4.TWin Roller[30] 105-10?K/s 10-100/m. strips. 5 8 5.Pendant 10 -10 K/s filaments. 25-125/jm. .
drop melt thick. extraction!36]
6.Melt drag[36,39] 102-104K/s. wire 0.1-8 mm. thick.
21
order of cooling rates and the resulting product of solidification
obtained in each of the process. The methods described till now are
used for the synthesis of foils tapes or filaments. Similarly there are
various powder production processes two of which are illustrated in
Fig.8[40-43]. Table 5 in a similar fashion lists the processes and
compares the cooling rates obtained in each of the processes.
2.2 III-V Semiconductors
The III-V semiconductors are formed by combining atoms from
group IIIA and group VA. These crystalline semiconducting compounds
are not alloys but definite chemical compounds with a 1:1 atomic ratio
between the III and V atoms which occupy alternate sites in the crystal
lattice. Table 6 lists some of the important properties of the III-V
compounds[44]. Nearly all the III-V compounds crystallize
into two forms of crystal structures: "Zinc blende" which is cubic; and
"wurtzite" which is hexagonal.[45]
2.3 Ternary III-V Antimonides
Combinations of the six IIIA and VA elements give the
antimonide compounds listed in Table 7. Ternary solids of the
antimonides can be prepared by melting corresponding binaries of these
antimonides; since they are mutually soluble in all proportions with
the exception of GaPSb and InPSb. [46] For most of the ternary
antimonides, a solubility rule derived by Foster and Woods[l] is valid
which states that complete miscibility should be expected in all of the
INDUCTION COIL
MELT
PRESSURE
l I I a • III i • j | I » • f't "1-J i •I'trr
CRUCIBLE
SPINNING CUP
TO DRIVE MECHANISM
1 k V V \ V V \ A
ORIFICE CHUOBUE
UOUIO / SOLID INTERFACE
MOLTEN DROPLETS
Figure 8. Powder preparation techniques: (a) Schematic arrangement of Rapid Spinning Cup (RSC); (b) Schematic arrangement of Twin Roller Atomization (TRA).[43]
M IO
Table 5. Powder Processing Techniques
Process Cooling Rate Particle size
1.Centrifugal 104-10*K/S
Atcmization[ 41 ]
2.Spark Eroslan[42] 10#-10#K/s 5-15/on
3.Supersonic gas
Atcmlzatlon[43] >10*K/S <20/m
4.Rapid Spinning >10eK/s <50jum
Cup[43]
5.TWln Roll 10tt-10#K/s <30fzm
Atcnlzaticn[ 43 ]
6.Plasma Spray 107K/S -
Deposltlon[43]
Table 6. Selected properties of III-V compounds, germanium and silicon [90].
Band Gap at Mdttef Point Vapor Pmui Material R.T. (ev) •o at M P. (aim) Cryul Strucnm
Si I.I 1 1420 10"* Cubic (diamond) Gc 0.67 936 I0"» Cubic (diamond) IN IO(Calc)»
— High Cubic (linc-bJcf«tc)
IN 4.6 (Calc)» >27001 High Ikugonal (wurohe) AIN >V >2900 >» Heaagtmal (wurtlite) CaN 3.2V >17001 >200* Hexagonal (wurtzite) InN U(Caic)> High Hcaagonal (wurtsite) BP i.V >ixxy >24» Cubic (linc-blcnde) AIP 2.42* >2000* High Cubic (mc-bknde) GaP 2.23 1467* »• Cubic (tinc-blcnde) InP IJJ 1060 23* Cubic (ainc-blcnde) BAa U(Cak)> >iooy High Cubic (ainc-blende) AiAi 116 >1700* High Cubic (tine-blende) CaAa 1.40 1231 09 Cubic (zinc-blende) InAl 0J3 Ml ass Cubic (nnc.btende) AlSb 1.74* ioso> <0.02 Cubic (awe-blende) CaSb 0.10* 706* <«x io-« Cubic (line blende) InSb 0.17 323 <IO*» Cubic (ainc-bhnde) InBi MctalKe 107 Vvrbo Ihnnona TIBi Metallic 230 Mnrhw Cubic oa
Table 7. Possible combinations of the III-V Antlmonldes [1].
BINARIES! TERNARIES QUATERNARIES
AIIIBVA"W aiiibii:cv aiiIbIIiciii dv aiiibw ww A ° y 1-y x 1-* x y 1-x-y x y 1-x-y x 1-x y 1-y
AlFSb AlGaPSb*
AlSb AlAsSb AlGaSb AlGalnSb AlPAsSb* AlGaAsSb
udSb GaPSb InGaSb GaPAsSb AllnPSb*
InSb GaAsSb AllnSb InPAaSb AlZnASSb
InPSb GalnFSb
InAaSb GalnFSb
Systems which have not been prepared yet.
26
systems where the lattice parameter difference is less than about 7.595.
However there are some exceptions as shown in Table 8.
2.4 III-V Ternary phase diagrams
Ternary III-V phase diagrams have three properties,
namely:
1) They are very useful in the LPE growth of III-V allays.
2) They are simple, and the liquid and solid solutions are
described well by simple solution models[47]. The simplicity of the
phase diagram is highlighted by the fact that the solid phase is
pseudobinary; that is, it has only one compositional degree of freedom,
and 3) The liquidus surface is simple. One liquidus surface, in
equilibrium with the pseudobinary solid, covers virtually the entire
ternary phase diagram.
The In-Ga-Sb ternary phase diagram has been determined
experimentally and has been verified by calculation. Further details
with regard to the calculation and evaluation of the phase diagram can
be obtained from references [48-59]. The usual ternary phase diagram is
a projection of the prismatic diagram upon a triangular base of the
prism. The liquidus isotherms are the intersection of the liquidus
surface and the isothermal planes. The solidus can be gained by a
similar way. The vertical plane intersecting at the two stoichiometric
binary compositions is called as "pseudo-binary" phase diagram. It
represents equilibrium between the stoichiometric liquid and solid. The
Table 8. Lattice parameters of binary constituents in III-V ternary antimonldes [1].
Binary pair Lattice parameter difference (nn)
(*)
AlSb-GaSb 0,00 UO ;0,65
AlSb-InSb 0,03itU 5^5
GaSb-InSb 0.038U 6,11
InAs-InSb 0.0U21 6,72
GaAs-GaSb 0,0W»2 7,52
AlAa-AlSb 0.0U7U 8,02
IaP-InSb 0,0611 9,88
GaP-GaSb 0.06UU 11,15
AXP-AISb 0,0673 11,65
28
"tie-lines" connect a point on the liquidus surface with the point on
the solidus surface with which it is in equilibrium. In the pseudo-
binary phase diagram the tie line connects the pseudo-binary liquidus
to the pseudo-binary solidus. In real cases the three III-V binary
compounds form an isomorphous system consisting of a liquid solution
(melt) and a solid sol\*tion. These two single phase regions are
separated by a three dimensional two phase region, which is bounded by
the liquidus surface at the top and the solidus surface at the bottom.
The three binary phase diagrams are the side - faces of the prism. Most
of the III-V ternary compounds have liquidus and solidus surfaces
characteristic of ideal systems.
The InSb-GaSb pseudo-binary phase diagram was determined by
Blom and Plaskett [49] as a special case from the ternary system where
the antimony fraction in the liquid remains constant at 0.5 and
therefore in the solid also. However for these calculations they found
that the pseudo-binary liquid is independent of the interaction between
In and Sb and Ga and Sb. Therefore, in general, the pseudo-binary
liquid can in first order be described by the interaction between the
two metals and is independent of the group V element. The pseudo-binary
phase diagram for InSb-GaSb system is shown in Fig.9.
2.5 Synthesis Of Bulk Crystals Of Ir fiaj Sb:
Primary investigations in this system were reported by Russian
workers Koster and Thoma [60] and Goryunova et al. [61] wherein they
900
InSb GoSb
Figure 9. The InSb-GaSb pseudoblnary phase diagram showing O :experimental data;*: differential thermal analysis; X ordinary thermal analysis; A : x-ray measurements; Q : [49]
30
tried to obtain single phase alloys as a result of progressive
annealing of powdered alloys at high temperatures. They tried to
achieve this with several compositions in this system. In later work
Woolley and Lees [62,81] showed that complete solid solution can be
obtained in the GaSb-InSb system provided that the alloys are annealed
for a sufficiently long period of time at the correct temperature. The
procedure followed was sealing the required quantities of the
appropriate compounds in evacuated sealed quartz tubes, heating to a
temperature above the melting point of the higher melting point
compound, and quenching in water. The resultant alloy was then powdered
and X-rayed following which the alloy would be again compressed and
resealed in vacuo and annealed to a temperature corresponding to the
melting point of the lower melting point component initially and then
progressively annealing it at higher temperatures depending on the X-
ray analysis until equilibrium was reached. However exceedingly slow
diffusion rates in the materials required long annealing periods
extending upto several weeks and in extreme cases several months before
the system showed any movement towards an equilibrium condition. The
problem was more severe in the case of solids as compared to the
powders of the same composition. The sluggishness of the system is
attributed to the very slow diffusion rates of the individual elements
in the solid state. Boltaks [63] studied the diffusion of these
elements in detail.
31
Ivanov-Omskii and Kolomiets [64] on realizing the difficulties
faced in the preparation of equilibrium solid solutions of these
compounds, used zone melting to achieve equilibrium conditions. Zone
levelling was also reported by McGroddy et al.[65] and Blakemore[66].
McGroddy et al.[65] reported that the ingots were polycrystalline with
grain size too small to allow preparation of single crystal samples.
Plaskett and Woods[67] reported the preparation of
polycrystalline alloy ingots - covering the entire composition range by
zone levelling techniques. They used zone travel rates of the order of
1 mm/hour. They examined the alloy metallographically, and obtained
lattice parameter values using X-ray diffraction. They also used
optical transmission techniques to determine the band edge. Joullie et
al.[56] reported the preparation of n-type single crystals of ternary
In aj.jjSb with 0<x<0.45 by means of a solution growth method (TGZM)
using indium as solvent and indium antimonide as seed. The charge
sandwich constituted polycrystalline Ga n j b obtained from a
Bridgman ingot. TGZM growing method consists in setting a seed charge
sandwich in a temperature gradient of the order of 80-
140°C/cm[68,69,70]. The charge which is dissolved, diffuses across the
liquid and crystallizes on the seed, while the solvent zone travels
towards highest temperat\ire regions. Growth and characterization of
solid solutions by TGZM was also performed by Hamaker et al.[68].
They obtained single crystal layers of upto 4 mm in thickness with
cross sections of 0.7 x 0.7 cm using <111> seed orientations of InSb at
32
a growth temperature of 420°C. Similar studies using TGZM was also
carried out by Michel et al.[71].
Solute diffusion growth for the preparation of ternary In Ga- ..
xSb crystals was reported by Miller et al.[72]. In this method a metal
alloy of a selected composition along with a quantity of the metalloid
element is placed in an evacuated sealed silica growth ampoule. The
growth ampoule is then placed in a growth furnace when the metalloid
will slowly distill or sublime and will enter the metal solution by
vapor transport at a rate that can be controlled by controlling the
metalloid reservoir temperature, Tc and its relationship to the
solution temperatures. Rodot et al.[73] also applied this method for
preparing sound ingots of Ga In .j Sb.
The travelling heater method (THM) of crystal growth, whereby a
solvent zone is caused to move through a solid feed material by the
movement of a heater; and first conceived by Pfann[69] in 1955 was
applied by Yip et al.[74,75]. They prepared cylindrical feed rods of
GaxIn2_xSb by gradient freeze technique in a vertical furnace[79].
Directional solidification of InxGa1_xSb was studied by Wilcox et
al.[76,77], They prepared ingots on the earth and in skylab, and
compared the influence of gravity on the crystal defect formation.
Joullie et al.[78] have also reported use of the vertical freezing
technique for preparing bulk GaxIn1_xSb using travel rates of the order
of 0.25-1 mm/hour. J. F. Yee [79] and R. A. Lefever[80]have also
reported preparation of bulk crystals of GaxIn1_xSb using the
travelling heater method.
33
Bachmann et al.[82] have succeeded in growing radially
homogeneous bulk single crystals using the czochralski method. They
synthesized single crystals of GajjIn -j Sb in the range 0.9<x<l.
2.6 Thin Film Methods
Crystalline and amorphous films of In Gaj Sb have been made by
several techniques such as liquid phase epitaxy, gas phase epitaxy,
vapor phase epitaxy, molecular beam epitaxy, flash evaporation and
multitarget sputtering. Greene et al.[83] have reported the growth of
single crystal epitaxial InxGa1_xSb thin films at compositions ranging
from x = 0 to x = 1, on a variety of substrates (BaF2, NaCl and InSb)
using multitarget sputtering.
Yano et al.[84] have reported growth of In Gaj.j Sb by
molecular beam epitaxy. They identified the relation of the arrival
rates of In and Ga molecules to the composition of the epilayer. The
films were grown on (100) surface of Cr doped semi-insulating GaAs.
Gotoh et al.[85] studied the growth of In0#2Gao.8sb on Cr doped
(100) GaAs using H2S gas for sulfur doping to change the conductivity
type.
Single phase InxGa1_xSb films by flash evaporation have been
prepared by Syamalamba et al.[86,87]. They prepared the films of
several compositions by controlling the deposition parameters and
characterized them for structural, compositional and optical
properties. Amemiya et al.[88] reported electrical property
34
measurements on films of ternary In Gaj Sb grown by a source
temperature programmed vacuum evaporation technique which is a
modification of a simple two temperature evaporation technique.
Wickersham et al.[89] carried out impulse stimulated "explosive"
crystallization of sputter deposited amorphous films of (In,Ga)Sb.
CHAPTER 3
EXPERIMENTAL SET UP AND PROCEDURE
3.1 MATERIALS PREPARATION
Preliminary experiments
Preliminary experimental work was performed by weighing
stoichiometric amount of InSb and GaSb required for obtaining the
desired composition in quartz tubes. The tubes were of standard wall
thickness and 0.3 cm. inner diameter. The tubes were batched with
99.999% InSb and 99.99% GaSb obtained from a commercial source1. The
samples were obtained in the form of 0.3 cm. pieces which were ground
to a powder in order to seal them in the tubes and batched in a glove
box back filled with nitrogen gas to prevent any atmospheric
contamination. The quartz tube was batched with InSb at the bottom
followed by GaSb, as shown schematically in Fig.10. The tubes were then
sealed after evacuating in a vacuum of 10~5 torr and heated in a
resistance heated vertical tube furnace to a temperature of 900°C.
The furnace temperature was measured with a thermocouple set
inside the furnace. The tubes were then held at that temperature for
1. Cerac Chemicals, Box 1178, Milwaukee, Wisconsin 53201.
35
Fused Silica Tube
Hot Zone
A
^
tO"5 Torr
In £b
— Go. Sb
Figure 10. Schematic diagram of synthesis of In Ga, Alloys. 1
37
eight hours to ensure good homogeneity and then quenched into a bath
comprising of a mixture of ethylene glycol and dry ice giving a final
temperature of -31°C. The quenched ingot gave sharp crystalline X-ray
diffraction peaks and hence it was decided to pursue other rapid
solidification techniques to obtain higher quenching conditions.
Initial test run-was conducted on a melt spinning set up which
.consisted of melting the alloy in a graphite susceptor placed in a
quartz tube. Ejection of the melt onto the wheel resulted in the
production of flakes. In order to collect these flakes, it was decided
to adopt the spinning cup mode of rapid solidification.
Design Of The Spinning Cup
The spinning cup basically comprised of two parts. An outer cup
made of aluminum, 17 cm. in diameter, and a disk made of copper, 0.3
cm. in thickness and 15.3 cm. in diameter. The copper disk was placed
at the bottom on the inside of the aluminum cup. The entire set up was
6.5 cm. in height and the aluminum cup was provided with a 5 cm. hub
which was fixed onto the shaft of a AC/DC motor capable of 10000 rpm at
0.5 HP with the help of set screws. In this way the spinning cup was
fixed to the motor. The motor speed was varied with the help of a
variac. This set up along with the motor was then placed inside a box
made of plexiglass covered on all four sides except at the top. This
arrangement was then placed close to the resistance heated vertical
tube furnace used earlier for tube quenching. Fig. 11 shows a schematic
38
Argon flow
Quartz tube ~ 10x12 mm
Copper disk Vfe in.
.V2 HP motor 10,000 RPM max.
Figure 11. Shanatlc of the 1RSP set up.
39
of the set up described above.
Melting and Ejection
The melting and ejection step involved melting the prealloyed
ingots of the desired composition. These ingots were made as described
earlier in the tube quenching experiments except that instead of
quenching; the tubes were allowed to cool in the tube furnace. The
tubes used in this spinning cup arrangement were 17 cm. long and were
provided with a nozzle. Two nozzle specifications were tried out in the
experiments, one with a diameter of 0.1 cm. and the other with a
diameter of 0.15 cm. Hie top end of the quartz tube was connected to a
solenoid valve by copper and tygon tubing. The solenoid valve was then
connected to the argon gas tank. The desired pressure of the argon gas
was obtained by adjusting the pressure in the regulator.
The melting sequence consisted of heating the furnace to the
desired temperature of 700°C or 900°C depending on the superheat
requirements. The quartz tube with the sample was then inserted into
the furnace to a depth of 10 cm. from the top of the furnace. Prior to
the introduction of the tube into the furnace a high flow of argon was
passed through the tube to remove the trapped air inside. The tube was
held in the furnace at the required temperature for a period of 30
minutes. Meanwhile the cup was fixed onto the motor and allowed to spin
at the desired rpm. The tube was then removed and the liquid metal
immediately ejected onto the rapidly rotating disk inside the cup, with
40
the help of the solenoid valve.
A large number of trials and experiments were carried out with
several compositions of the alloys at various rpm settings of the motor
using tubes provided with nozzle diameters of 0.1 and 0.15 cm. The rpm
of the cup was calibrated to the variac setting using a stroboscopic
light. In this way the cup was made to rotate between 4000-10000 rpm.
The flakes and powders were analyzed using X-ray diffraction, optical
microscopy, SEM, TEM and Auger electron spectroscopy for characterizing
the microstructure. Phase transformation investigations were conducted
using the DSC.
The flakes and powders were examined as is for topographical analysis
and etched to reveal the microstructures. The surface as well as the
cross section of the powders and flakes were examined.
Heat Treatment
The as-prepared flakes were sealed under a vacuum in quartz
tubes and heat treated in a drive in furnace to temperatures of 100,
200, 300 and 400°C for a period of two hours at each of the
temperatures to induce annealing which are discussed in the later
chapter. The samples were then again mounted in casting resin and
polished on which the electrical resistivity measurements were made.
3.2 MATERIALS CHARACTERIZATION
X-ray Diffraction
X-ray diffraction analysis was performed on the as-prepared
41
flakes and powders on a standard diffractometer using monochromatic
copper Ka radiation provided with a nickel filter. The beam slit used
was 1°medium resolution (MR) and the detector slit used was 0.2°MR. The
furnace cooled ingots were also subjected to X-ray analysis by crushing
to -325 mesh size fractions. The instrument was calibrated for the
proper 28 values using an internal standard of high purity silicon.
For all the runs a scanning rate of two degrees per minute was used.
Optical Microscopy
Optical microscopy was done on the flakes and powders which
were etched to see the grains and the columnar structures during the
solidification of the alloys. This was done by mounting the flakes in a
casting resin to see the morphology on the surface. In order to see the
cross section, the flakes and powders were sandwiched between the
casting resin and then cut at right angles to see the cross section of
the solidified flakes and powders.
Scanning and transmission electron microscopy
Scanning electron microscopy (SEM) and energy dispersive X-ray
analyses (EDAX) analyses were performed on the as-prepared powders and
flakes. The flakes and powders were mounted on standard sample stubs
and sputtered with Au/Pd. The etched samples were mounted in a casting
resin, and the entire resin surface excluding the sample was coated
with colloidal carbon to prevent charging of the electron beam.
42
Transmission electron microscopy(TEM) was performed on crushed
powdered samples dispersed in n-batanol and floated on carbon coated
copper grids. Selected area diffraction was used to determine the
amorphous or macrocrystalline nature of the alloys.
Scanning Auger Electron Microscopy
Scanning Auger electron microscopy was done on the as-prepared
flakes to see if there was any oxidation and to see if there was any
formation of oxides. This was done by taking area scans and depth
profiling by argon sputtering at 2keV using an electron beam of 3keV.
Thermal Analysis
Differential scanning calorimetry (DSC) was performed on as-
prepared flakes and powders. This was done to see the effect of rapid
solidification on the structure and to perceive any phase
transformations if any. The analysis was performed in a Dupont 1090
Thermal Analysis System,2 which is computer interfaced allowing fast
real time data acquisition at maximum sensitivity.
The software automatically determines peak positions. The as-
prepared powders and flakes (typically weighing 10-20 mg) were sealed
in copper pans; a small opening was provided in the pans to prevent any
pressure build up in case of volatile elements being present in the
samples.
2. Dupont 1090 Thermal Analyzer, E.I. Dupont and deNemors Inc., Wilmington, Delaware.
43
The instrument could record any thermal changes up to a temperature of
600°C. A heating rate of 10°C/min was used for all the samples, and a
neutral atmosphere of nitrogen gas was maintained in the DSC cell.
Electrical Resistivity Measurements
A four point probe was used to measure the resistivity of the
heat treated flakes. The samples were annealed for two hours at
temperatures of 400,500,600 and 700°C, respectively, and then mounted
in casting resin and polished. Hie measurements were performed at room
temperature after making the necessary correction for the sample shape
and thickness. These corrections were obtained from reference[97].
CHAPTER 4
RESULTS AND DISCUSSION
4.1 X-ray Diffraction Analysis
The X-ray diffraction analyses were performed on all rapidly
solidified samples which are discussed separately in two subsections
entitled General Analysis and Specific Analysis.
General Analysis: This section includes the analysis performed on the
tube quenched ingots and the analysis on the resultant flakes and
powders obtained by melt spinning initial trial melts. The compositions
investigated and the melt spinning conditions are listed in Table 9.
Tube quenching experiments were performed on alloys of composition
In0.53Ga0.47Sb' In0.44Ga0.56Sb' In0.46Ga0.54Sb' and In0.41Ga0.59Sb"
X-ray diffraction analysis performed on all these samples
indicated that the samples were strongly crystalline, and it was thus
necessary to pursue the alloy processing by melt spinning to induce
microcrystallinity or amorphism. All the compositions listed in Table 9
were initially ejected onto the Cu disk inside the spinning cup(see
Fig.11) at a distance of 2.54 cm. from the center. At a relatively low
speed of 4000 rpm, flakes were obtained and at 6000 rpm, a combination
of flakes and powders were obtained as shown in Table 9. The X-ray
traces of the ingots for each of the above compositions were used to
44
Table 9. Melt Spinning Compositions with Melt Spinning Parameters
Composition T melt rpm Velocity Pressure Nozzle Ejection run Form Remarks
(°C) (m/s) x 104Pa diameter distance (cm) (cm)
.ls .e?813 960 6900+100 18.4 7 0.09 2.54 1 Flake -
In0.13Ga0.87Sb 967 10000 +100
27.26 7 0.09 2.54 2 Flake -
In0.16Ga0.84Sb 940 8000+100 21.3 8 0.09 2.54 3 Flake -
966 5600+100 14.62 7 0.09 2.54 4 Flake -
860 8000+100 21.3 2 0.09 2.54 5 Flake powder
-
700 9000+100 23.93 4 0.09 2.54 6 Flake powder
-
In0.51Qa0.49Slb 865 6500»+100 17.3 4 0.09 2.54 7 Flake -
835 9500+100 25.27 2 0.09 2.54 8 Flake powder
-
o.si o 813 820 9500+100 25.27 4 0.09 2.54 9 Flake powder
melt held for 3 sec.
Continued
Table 9. Melt Spinning Compositions with Melt Spinning Parameters
Composition T melt rpn Velocity Pressure Nozzle Ejection run Form Remarks
(°C) (m/s) x 104Pa diameter distance (an) (cm)
825 9500+100 25.27 4 0.09 2.54 10 Flake powder
melt held for 18 sec.
In0.54Ga0.46Sb 700 9500+100 25.27 2 0.09 2.54 11 Flake powder
cup fixed vertical
.ga o.o?813 700 10000 +100
26.6 4 0.09 2.54 12 Flake powder
-
710 9500+100 25.27 4 0.09 2.54 13 Flake powder
-
In0.446a0.56Sb 680 10000 +100
26.6 4 0.09 2.54 14 Flake powder
-
In0.44Ga0.56Sb 720 9500+100 25.27 2 0.09 2.54 15 Flake powder
-
In0.44Ga0.56Sb 720 8000+100 21.3 1 0.09 2.54 16 Flake powder
-
In0.44Ga0.56Sb 830 10000 +100
26.6 2 0.09 2.54 17 Flake powder
-
Continued
Table 9. Melt Spinning Compositions with Melt Spinning Parameters
Composition T melt rpn Velocity Pressure Nozzle Ejection run Form Remarks
(°C) (m/s) x 104Pa diameter distance (cm) (cm)
In0.44Ga0.56Sb 749 4250+100 33.91 4 0.15 2.54 18 Flake 3 sizes length : > 5mn.l8A = 5mm.18B < 5mm.18C
In0.44Ga0.56Sb 735 6500+100 51.86 4 0.15 2.54 19 Flake 19F powder 19P
In0.44Ga0.56Sb 735 8020+100 63.99 4 0.15 2.54 20 Flake 20F powder 20P
48
perform lattice parameter calculations, summarized in Table 10.
These calculations provide evidence that the alloys do not
yield an equilibrium solid solution upon solidification. In addition,
some ingot compositions show presence of two non- equilibrium phases
and on subsequent melt spinning there is either a retention of one of
these phases or a different non-equilibrium phase formation as
indicated in Table 10. In the others the non-equilibrium phase is
retained on quenching. A common feature however is a noticeable
disappearance of peaks in the traces obtained for the flakes and
powders indicating disordering as shown in Fig.12 which is a
representative trace for alloys of composition Ino.44Gao.56s 3, A so the
X-ray traces for the powders are relatively broad in comparison with
those obtained for the flakes; this is a result of smaller crystallite
sizes obtained in the case of the powders, because cooling rates in the
powders were probably greater than the cooling rates in the flakes.
These data suggest that the powder is probably formed in the present
processing by a mechanism similar to that described by Glickstein[41]
wherein the liquid metal on contact with the disk disintegrates into
molten droplets which are solidified either in air or on impinging the
walls of the cup. This is the mechanism for powder formation in the
centrifugal atomization technique described in reference[ 41 ].
Because of the large surface area associated with these
particles there is rapid cooling resulting in high cooling rates. The
broadening of the peaks which is induced due to reduction in
Table 10. Lattice Parameter Values
Composition Lattice Parameter A°
Literature value [44]
Ingot Powder Flake Structure Literature value [44] a2
.IS .S?813 6.15 6.095 - - 6.095 diamond
lh0.1GPa0.84sb 6.16 6.095 - 6.095 diamond
In0.43Ga0.57Sb 6.265 6.447 6.29 6.23 6.23 diamond
In0.44Ga0.54Sb 6.27 6.095 6.29
'6.095
6.29
6.095
diamond
6.275 6.447 6.29 6.192 - diamond
In0.51 0.49s*5 6.295 6.447 6.29 6.29
6.095
6.29 diamond
In0. . 07Sb 6.48 6.447 6.447 6.447 diamond
Powder
(III) (220)
(311) a •f/AN—, ,J T-+-I : 1—I—I
Flake (/)
w-O
(220)
C
Ingot
(220)
(311) (422) uu (331)
50 48 ANGLE (20 degrees)
Figure 12. X-ray trace of ingot, flake and powder obtained from melt of run 19.
8
51
crystallite size was observed only in powders obtained from melts of
initial composition InQ 51GaQ 49Sb and In0 44Ga0 56Sb. This suggests
the marked influence of RSP only on these two compositions. X-ray
traces for powders obtained for other compositions showed sharp
crystalline peaks further supporting the above inference.
Specific Analysis
Specific X-ray diffraction analyses were conducted on flakes
arc! powders obtained for alloys of initial composition Ino.5iGao.49Sb
and In0-44Ga0 56Sb.
Based on previous work of others [ 2 ], experiments were designed
to assess the effect of undercooling on the structure by reducing the
superheat. The experiments were carried out by removing the quartz
tube from the furnace after the alloy was melted and homogenized for a
half hour at 800°C (a superheat of 160-200°C) and holding the tube for
several seconds in order to decrease the temperature before ejecting
the alloy. Three experiments based on this procedure were conducted on
the Ino.5iGao.49st> composition. The experimental parameters are listed
in Table 9. They have been identified by runs 7-10. All these runs
resulted in the production of flakes and powders which exhibited
similar X-ray traces. Lattice parameter values for these alloys are
shown in Table 10. Again it is clear that there is formation of mixed
phases and a retention of one of the phases in the flakes and powders
along with a marked broadening effect of the peaks for the powders.
52
In all the above experiments the parameters varied In melt
spinning were the holding times, temperature and the disk velocity.
From the X-ray analysis it was concluded that the velocity of the disk
and the redaction in superheat played a significant role in optimizing
the effects of rapid quenching. Therefore melt spinning of
In0.44Ga0.56sb was performed at a reduced but constant superheat of
100°C while only varying the disk velocity. Further, all the
experiments on this alloy were performed by ejecting the liquid metal
at the periphery of the copper disk to obtain maximum peripheral
velocity to act on the metal so that along with the centrifugal force
effect there would also be shearing of the metal droplets inducing more
effective quenching conditions.
These melts are identified by runs 18, 19, and 20. Run 18
resulted in big flakes ( > 5 mm length), small flakes (=< 5mm length)
and fines (< 5mm length ) identified as 18A,18B and 18C, respectively,
in the discussion below, whereas runs 19 and 20 resulted in the
production of flakes (identified as 19F,20F) and powders (identified as
19P,20P). The X-ray traces for all these melts showed essentially
similar peaks to those obtained for Ir\).44Qa0.56s'3 sh°wn in Fig. 12 for
run 13. Runs 13 and 19 showed identical X-ray traces. The X-ray traces
however showed a relatively larger amount of peak broadening in
comparison to the previously mentioned compositions. This was because
the experiments were performed by ejecting the liquid metal on the
periphery of the disc.
53
In order to see the effect of heat treatment on the rapidly
solidified flakes, X-ray analyses were performed on the flakes of
composition In0.5iGa0<49Sb obtained from run 7, after they were
subjected to an annealing treatment at temperatures of 100, 200, 300
and 400°C for a constant period of two hours. The X-ray trace obtained
for the flakes annealed at 400°C was identical to the trace obtained
for the ingot. In the cases of the samples annealed at temperatures
0 other than 400 C, the traces indicated a gradual increase in the
intensities and the reappearance of peaks, which were absent in the
quenched flakes.
In summary, it is reasonable to conclude that the flakes and
powders contain a mixture of metastable phases, similar to the
unquenched ingot, and show a general broadening of peaks, which
suggests very fine grains due to rapid solidification. Flakes obtained
from run 7 retained one of the phases seen in the unquenched ingot
(Table 10); however heat treatment of these flakes results in the
reappearance of the original phases present in the initial ingot.
4.2 Surface Morphology
The as-prepared flakes and powders were examined under the SEM
using EDAX to observe the morphology of the flakes. Fig. 13 shows
micrographs of surfaces of the flakes away from the copper substrate.
It appears that liquid metal droplets have solidified on the surface
to give it a very irregular morphology. At high magnification these
Figure 13(a). SEM micrograph showing surface of flake obtained from run 9, mag. 33 X.
Figure 13(b). SEM micrograph showing surface of flake obtained from run 20, mag. 1500 X,
Figure 13(c). SEM micrograph showing precipitates in flakes obtained from run 1, mag, 4200 X.
56
solidified droplets appear like precipitates (see Fig.l3(c)).
SEM(EDAX) was performed on the precipitates and on the matrix
in order to study any variations in composition. The precipitates
showed a considerable decrease in the Ga counts in comparison to the
matrix.(see Fig. 14 and 15) Hence it can be concluded that the
precipitates are richer in InSb as compared to GaSb. These precipitates
were observed in all the compositions. However these precipitates were
only seen on the surface of the flakes solidified away from the copper
substrate. The surface of the flake adhering to the copper disk showed
a much smoother morphology in comparison to that discussed above. There
was also no compositional variation.
The powders were also examined voider the SEM. Fig.16 shows the
morphology of the powders in a series of micrographs taken on powders
obtained from run 8, 9, 10 shown in Table 9. It can be seen that the
powders are irregular with particle sizes ranging from 1 micron to
several 100 microns. A noticeable feature is the cluster of spherical
particles in photomicrograph 16(d) with particle sizes being 2-3
microns. Micrograph 16(e) shows the fractured part of a large powder
particle. The fractured surface shows the columnar growth indicating
the directional mode of solidification. Micrographs 16(f) and 16(g) on
the other hand do not show the formation of any spherical particles.
These were the powders obtained from run 8 and 9. The particles are
very irregular, and the sharp edges, in fact, indicate that they are
probably formed by large flakes fragmenting into smaller ones. However
2888
COUNTS
PRECIPITATE ANALYSIS
ENERCT (KEVJ
Figure 14. EDAX analysis of the precipitate observed on the flake surface.
2
IBM
MATRIX ANALYSIS
COUNTS
t.a te.e ENCRCY IKEVI
Figure 15. EDAX analysis of the matrix on the flake surface.
Figure 16(a). SEM micrograph showing the powder obtained from run 10, mag. 360 X.
Figure 16(b). SEM micrograph showing the powder obtained from run 10, mag. 3600 X.
SEM micrograph showing the powder obtained from run 10, mag. 1090 X.
SEM micrograph of powder obtained from run 10, mag. 1620 X.
61
Figure 16(e). SEM micrograph showing fractured powder particle obtained from run 10, mag, 1620 X,
Figure 16(f). SEM micrograph showing the powder obtained from run 9, mag. 660 X.
Figure 16(g). SEM micrograph showing the powder obtained from run 8, mag. 430 X.
*
63
the particles occur in various shapes so that they could be occuring by
a combination of 'centrifugal atomisation' and by fragmentation.
4.3 Thickness Measurements
Figure 17 shows micrographs of cross sections of a flake.
Micrograph 17(a) shows a part of the flake surface in contact with the
copper disk, while 17(b) shows the surface not exposed to the disk and
hence exhibits a very rough surface in contrast to the relatively
smooth surface shown in micrograph 17(a). However both micrographs
indicate the columnar mode of solidification which is expected since
heat flow is basically directed towards the copper disk. The striations
on the surface seen in Fig. 17(a) are due to the flake being in contact
with the disk.
The thickness of the flakes were measured from micrographs of
one or two random flakes flakes obtained from melt spinning each of the
various compositions listed in Table 9. Thickness measurements were
made by taking into account the distortion in the secondary electron
image due to the tilting of the specimen. The values, estimated after
incorporating the correction were then correlated with the tangential
speed of the disk at the point where the liquid metal was ejected. A
plot of the flake thickness against the speed is shown in Fig. 18. Each
point on the plot represents an average value of 8-10 thickness
measurements. The bar around each point is a measure of the standard
deviation. The plot shows that the flake thickness decreases with
Figure 17(a); SEM micrograph of cross section of flake obtained from run 1, mag. 1700 X,
Figure 17(b). SEM micrograph of cross section of flake obtained from run 19, mag. 770 X.
Ti 5~TT VELOCITY fm/s)
¥ lo So" lo 80
Figure 18. Variation of flake thickness with disk velocity.
o> CJl
66
increasing speed, but for speeds greater than 35m.s-1 a slight increase
in overall flake thickness is observed. This may suggest that some
flakes lose contact with the disk resulting in higher thickness
values. It may be noted that the plot may not be a true representation
of the entire process since the measurements were made on one or two
random flakes.
The thickness measurements were also used to estimate the
cooling rate after making a reasonable assumption that the heat
transfer coefficient at the interface between the flake and the
substrate is infinity, rendering conditions of ideal cooling to be
applicable. This assumption is applied from the fact that the majority
of the flakes exhibited excellent wettability!19]. Based on this
assumption, thecooling rate,
(CR) «=ci/d2 (2)
where CR = cooling rate from the initial splat temperature to the half
temperature, given by,
*1/2 = 1/2(TS+TB) (3)
where Ts = initial splat temperature and TB = initial substrate
temperature and d = thickness of the flake. Ruhl[19] made calculations
based on a one dimensional heat flow analysis and obtained a
proportionality constant of approximately 81 cm2°C/S which agreed quite
well with the analytical solutions obtained for the cooling rate from
Cars law and Jaegar[98]. Ruhl's analysis gave the following relation for
67
the cooling rate,
CR = 81/d2 (4)
This square relation is reported to hold true for ideal cooling
conditions and is found to be marginally affected by changes in
solidification temperature, initial substrate temperature, and splat
and substrate material. Assuming the validity of the above equation, a
range of cooling ra.tes were estimated for the flake thickness
experimentally determined. Fig. 19 shows the variation of the estimated
cooling rates with experimentally determined flake thickness. As can
be seen cooling rates varied from 3.1 x 106 to 3.2 x 10 °C/s,
the higher cooling rates being obtained for flakes of diminishing
thicknesses. It may also be noted that thickness is also involved in
the parameter hd/kg (d=flake thickness, h=heat transfer coefficient and
kg=thermal conductivity of the flake) which determines whether ideal
cooling exists. As *d' becomes smaller, often the cooling will cease to
be ideal unless ~h' is very large. In our discussion since the flakes
made very good contact with the substrate we assume ideal cooling
conditions whereby, then the above square law holds. [19]
4.4 M icrostructure Analysis
The microstructure of the flakes and the powder were recorded
by optical microscopy and scanning electron microscopy after etching
them with an etchant consisting of the composition 2HF:2HAc:3HN03:6H20.
Initial composition of the etchant was obtained from the
To Jo THICKNESS (MICRONS)
¥~ Jo 80
Figure19. Plot of experimentally determined flake thickness (d, microns) versus cooling rate (°K/s) calculated assuming the validity of the equation Cd =81 [19]
o» CD
69
literature[76], however the etchant was very strong and had to be
diluted with water until finally, etching the flakes with the above
composition for 1 second worked extremely well in revealing the grain
structure of the rapidly solidified flakes and powders. The etchant
however failed to reveal any structure on the furnace cooled ingots.
The ingots on the other hand being bireflectant [ 79 ] show the grains by
reflection of light incident on a flat cross section of the ingot at an
angle. Fig.20(a) shows the micrograph of the cross section of the,
ingot with the grains being revealed by the property of bireflectance
exhibited by this alloy.
The figure shows random orientation of the grains since the
ingot was just allowed to cool in the furnace. Moreover the grains are
not uniform, being elongated more in one direction than the other. An
average grain size of 970 microns was estimated for the ingot by taking
an average of 13 measurements for several grains seen in Fig. 20(a).
Fig. 20(b) shows optical micrograph of the surface of the flake etched
with the above etchant. The flake was obtained by rapidly solidifying
the ingot of initial composition Ino.51GaQ.4gSb,represented by run 7 in
Table 9, the surface of which clearly shows the cellular grain
morphology.
SEM micrograph of the same area is shown in Fig. 20(c). The
pores seen in the micrograph correspond to the etch pits. Since the
topographical analysis of the flakes revealed presence of second phase
precipitates on the surface the etched surfaces were examined for any
Figure 20(a). Optical micrograph showing the cross section of an ingot of composition Ino.46Gao,54Sb» ma9- 10
Fiqure 20(b). Optical micrograph showing the etched surface of the flake obtained from run 7, mag. 1000 X.
SEM micrograph showing the etched surface of the flake observed in Fig. 20(b), mag. 2300 X,
72
compositional inhomogeneities. A number of spectra were taken on the
grains and on the grain boundary to observe any changes. However no
differences in the counts for any of the components was seen; this
means that the precipitates formed only on the surface and subsequent
polishing during sample preparation eliminated them.
The cross section of the flakes and powders were then examined
to observe the structural morphology. Micrographs in Fig. 21 show the
etched cross section of the flakes. Fig.21 (a) shows the optical
micrograph of the flake of composition Ing gjGaQ 49Sb, obtained from
run 7 in Table 9, whereas Figs. 21(b),21(c) show the microstructure of
etched cross section of flakes of composition Ino.44Ga0.56sb obtained
from run 14 in Table 9. The micrographs clearly show the columnar
grains traversing from the top to the bottom which is the direction of
heat flow towards the copper disk.
Micrographs 21 (a),21(b), show the solidification morphology and
in some cases clusters of very fine dendrites. Fig.21(c) represents an
SEM micrograph of a section of Fig.21(b). The micrograph shows the
branching of the dendrites. The photomicrograph shows an irregular
morphology of the dendrites which may be because of the rapid rate of
heat extraction induced by the process. Another reason is the rotation
of the disk which causes shearing of the liquid metal.
There was no compositional variation seen in the cross section
of the flakes. The cross section of the powders also show similar
behaviour as the flakes. Figs. 22(a) and 22(b) show micrographs
Figure 21(a). Optical micrograph showing the cross section of flake obtained from run 7 etched with the modified reagent, mag. 1000 X.
: * V* ' •
w <4 V-cm; f s -"•? * —
BR 3
• ,*w«. .. - •
' • J
-A v?**-« %
Figure 21(b). Optical micrograph showing the cross section of flake obtained from run 14, mag. 1000 X.
SEM micrograph showing the cross section of flake observed in Fig. 21(b), mag. 1700 X.
Figure 22(a), SEM micrograph showing the etched cross section of powder obtained from run 14, mag. 3700 X.
Figure 22(b). Optical micrograph showing the etched cross section of powder obtained from run 14, mag. 1000 X.
76
t
Figure 22(c). Optical micrograph showing the etched cross section of powder obtained from run 10, maq. 1000 X.
77
obtained from the cross section of powders of composition
InQ 44Ga0 5gSb obtained from run 14 in Table 9. The micrographs show
the columnar grains similar to those shown for the flakes. Fig.22(c)
shows the micrograph which represents the cross section of a powder of
composition Ino.5iGao.49sb represented by run 10 in Table 9. The
micrograph also shows columnar grains except that certain parts of the
powder remain unetched suggesting the presence of amorphous regions.
Similar evidence of unetched amorphous regions was also observed in
splat quenched Fe80B2oflakes investigated by Turnbull[25].
In order to see the effect of the cooling rate and disk vlocity
on the microstructure, flakes obtained from run 7, 18, 19, and 20 were
etched and examined under the optical microscope and the SEM. Figures
23 and 24 show the optical and SEM micrographs respectively for the
flakes obtained from runs 18,19 and 20. Grain size estimates were made
from each of the micrographs using the straight line method described
by Underwood et al.[91]. A plot of the grain size against the disk
velocity is shown in Fig.25. Hie plot exhibits a behaviour similar
to the trend seen in the thickness - velocity plot in Fig. 18. The
grain sizes decrease with increasing disk velocities but increase at
very high velocities due to insufficient disk-flake contact at such
high velocities. Therefore, an important conclusion from Fig. 18 is
that the data reflect a velocity band in the range 45-60 m/s where the
effect of rapid quenching is at its peak. This also suggests that glass
formation tendency is the highest in this velocity band for the
Figure 23(a). Optical micrograph showing the etched surface of the flake obtained from run 18, mag. 1000 X.
Figure 23(b). Optical micrograph showing etched surface of the flake obtained from run 20, mag. 1000 X.
Optical micrograph showing the etched surface of the flake obtained from run 19, mag, 1000 X.
SEM micrograph of etched surface of flake observed in Figure 23(a), mag. 3100 X.
grain size ingot 970 licrons .
7-cn z o cr o 6 -l-H
LU IVJ I—I CJ1
cr CD
5 30 35, , 40 VELOCITY (m/sl
Figure 25. Variation of grain size with disk velocity.
s
I
81
particular alloy composition { Ino.44Gao.56Sb' 1,1111 19 '
Figures 26 and 27 show the bright field TEM image and an
electron diffraction pattern of the same area obtained on a flake
from run 13. The diffraction pattern shows the presence of rings with
spots indicating the flakes to be microcrystalline. The presence of
spots is because of the random orientation of the small crystallites
which intersect the electron beam producing a diffraction pattern with
a series of spots. The bright field image in Fig. 26 shows presence of
several phases. Fig.28 is a diffraction pattern of powder from run 20
(20P). The diffraction pattern shows rings with an amorphous halo,
which indicates the presence of a small fraction of amorphous phase.
However exposure of the sample to the electron beam during microscopy
caused crystallization indicating that the alloy was highly beam
sensitive. Evidence of this is seen in Fig.30 which is the diffraction
pattern of the same area after exposing the sample to the beam for ten
minutes. A distinct sharpness in the rings is seen which indicates
microcrystallinity in the sample. The rings in the diffraction patterns
correspond to a diamond cubic structure. Bright field image of the
same area around the halo in Fig. 28 showed the formation of very small
crystallites of the order of 32 nm. The bright field image is shown in
Fig.29.
Dark field image of the same area obtained by tilting the beam
is shown in Fig. 31. The image shows the tiny crystallites, however
the streaking in the image is because of the instrument drift induced
t 0-4 Mm |
Figure 26. Bright field image of flake obtained from run 13, mag. 80,000 X.
f
s
I
Figure 27. Diffraction pattern obtained from an area of flake observed in Figure 26.
8 3
Figure 28. Diffraction pattern obtained for powder obtained from run 20.
Figure 29. Bright field image of the same region from which diffraction pattern in Figure 28 was obtained, mag. 31,200 X.
8 4
Figure 30. Diffraction pattern obtained for powder obtained from run 20 after exposing the sample to the electron beam for 10 minutes.
, - > ' * ** Y s*. 1
* • / fy S' />;'
/ ;
Figure 31. Dark field image of the region from which diffraction pattern in Figure 28 was obtained, mag. 31,200 X.
85
in the sample. This has become very prominent due to the very long
exposure of hundred seconds introduced during photography of the image.
EDS spectra of these areas were obtained to confirm that the patterns
were originating from the sample and not from the carbon film on the
grid. The spectra confirmed the presence of In, Ga and Sb.
Auger Electron Spectroscopy performed on the surface of a flake
( obtained from run 14) shows the presence of In, Ga, Sb, 0,C and Cu.
The presence of Cu on the suface is possibly due to quenching on the
copper disk. Fig.32 shows the Auger electron spectra indicating the
presence of the individual elements. Depth profiling was performed to
determine if oxide formation occured in the processing of these alloys.
The profiling was obtained by argon ion beam etching while continuously
monitoring the elemental Auger lines. The ion gun energy used was 2 keV
and the electron beam vised was 3 keV. The depth profiles are shown in
Fig.33 which do not show any substantial amounts of 0, C and Cu in
the sample for a sputtering time of ten minutes.
4.5 Transformations
In order to detect the presence of any metastable phases DSC
analyses were conducted on the ingot and melt spun samples of
composition Ino.44Gao.56s obtained from run 18,19 and 20. The
analyses were conducted on the powders as well as the flakes. These DSC
traces for the ingot and the flakes are shown in Fig.34. The DSC trace
for the ingot shows an exotherm at = 300°C which confirms the non-
86
• •
•
• •
; Gfl C- IN-h
: N 1
SB
cu G fl
>
1111 i
:
: 0 l
:
i
180 200 300 400 S00 600 700 800 900 1000 1100 1200 KINETIC ENERGY, EV
Figure 32. Auger electron spectra showing presence of C, 02, Cu, In, Ga, Sb in flake obtained from run 16.
87
J0
a8
6
4
2
0
J0
a8
6
4
2
0
. . :
:: : A :: :
A
UUl A \ I
: n. \ f / \ :
: n. V V J 1 : n.
3 4 5 6 7 8 SPUTTER TIHE (HIH.)
10
T : • . •
; ;
J ;
i
: • : •
1 c 1 i
: •
l i l l l 2 3 4 5 6 7
SPUTTER TIHE (HIH.) 10
i i i i t i i i i i i i i i M i i i i i i i »i i ii ii i
Or 1
2 3 4 5 6 7 8 9 1 0 SPUTTER TIHE (HIN.)
Figure 33. Depth profiles for Cu, C and elemental 0 in flake obtained from run 16.
88
Ingot
t -a}
i
t -ioj
i-0 lea 2 ee
f: »• 300 400 S00 800 7£
Tump«raiur<i C#C5
Figure 34. DSC traces of ingot and flakes obtained from run 18,19 and 20.
89
equilibrium nature of the ingot indicated by the X-ray results. The X-
ray results for the ingot as described earlier indicated presence of
non-equilibrium phases confirmed by the lattice parameter calculations.
Similar exotherms were seen for the flakes melt spun at low and
high rpms corresponding to run 18 and 20 identified as 18A and 20F.
This indicates that the flakes melt spun at relatively low and high
rpm's retain metastable phases identical to the ingot. But the DSC
trace for the flakes obtained from run 19 identified as 19F does not
show the exothermic peak. Hie DSC traces for the corresponding powders
are shown in Fig.35 which show a similar trend to those obtained for
the flakes. The only difference is that the trace identified as 18B and
18C which corresponds to the fines obtained from run 18 in addition to
the powder obtained from run 19 does not exhibit the exothermic peak.
This observation supported by the grain size-velocity plot (see
Fig.25) suggests the possibility of an amorphous phase being formed in
this sample although the X-ray traces were similar to those obtained in
the case of the powders and flakes obtained from runs 18 and 20. (see
Fig. 12). Also, the fraction of the crystalline phase may be
insufficient for the DSC to register the exothemic peak characteristic
to the phase. A summary of these conclusions are shown in Table 11.
Thermal analyses were also conducted on the powders obtained
from run 8,9,10 for samples of composition Ino.51GaO 4gSb. The DSC
traces for these runs are shown in Fig.36. There is no evidence of any
exothermic transition, however the trace for run 10 where the melt was
90
0->
20 P
18C
-3
-4-
B 0 U.
18 B 0 1 -3"
- j . .
19 P
-3-
- S -
IBS
Figure 35. DSC traces of corresponding powders obtcdned from runs 18, 19 and 20.
Table 11. Summary of XRD and DSC Analysis
Sample Composition IriQ. 44Ga0.56s*3
Form Run $
Phase Present (XRD)
Phase present (DSC)
Flake 18A A A
Flake 18B A -
Powder 18C A -
Powder 19P A -
Flake 19F A
Flake 20F A A
Powder 20P A A
Ingot A A
* The letter 'A' represents the metastable phase
92
1 0 -
0-
-10..
-10
0
-10
» -20 o u. .*> s -30 X
-40-
-50. .
-ea
run 10
run 9
run 8
100 200 300 400 500 600 700 T«npTatur« <®D
Figure 36. DSC trace of powder obtained from runs 8, 9, and 10.
93
held for 18 seconds before ejecting the liquid metal shows a slight but
noticeable change in slope at approximately 400°C which indicates a
probable glass transition. This reasoning can be supported by the
observation of certain unetched regions in the microstructure of this
sample (see Fig. 22(c)). Apart from this observation all the traces
exhibit consistently identical endotherms.
4.6 Electrical Property Measurements
Bulk electrical resistivity measurements were made at room
temperature only on flakes of composition InQ GaQ gSb obtained from
run 7. The flakes were large enough to permit electrical resistivity
measurements using the four point probe. Flakes obtained by melt
spinning other compositions were too small to perform electrical
resistivity measurements. The measurements were made on flakes
annealed at temperatures of 100,200,300,and 400°C for a constant period
of two hours. The resistivity measurements were made by incorporating
corrections for the thickness and the shape of the flake. [97].
A plot of these values for the various temperatures is shown in
Fig. 37. Each resistivity value represents an average of twenty
measurements performed on each flake. The bar around each point is a
measure of standard deviation. The plot is unusual since one would
expect the resistivity to decrease with increasing annealing
temperatures due to stress relaxation. Resistivity values reported at
room temperature by earlier investigators for bulk ingots of similar
• unquenched ingot 17 .109-
.099-
.079-
.069-
.059-
.049-
.04-
.03-
.02-1
.01-
400 600 Too 300 ANNEALING TEMPERATURE (degree K)
200 800
Figure 37. Variation of bulk electrical resistivity with annealing temperatures for flakes obtained from run
95
compositions are shown in Table 12. The values obtained by them fall in
the vicinity of the values obtained in the present investigation for
annealing temperatures greater than 200°c.
An explanation for this behaviour is suggested: The rapidly
solidified flakes contain, in part, a metastable phase as inferred from
the X-ray analysis and different metastable phases can have different
resistivity values. Thus the rapidly quenched flake consisting of
metastable phases at room temperature has a resistivity value of 0.02
ohm-cm which is higher than that measured for the ingot. The higher
value of the quenched flake is due to the disorder induced in the
sample by the quenching process[96]. On subsequent annealing there is a
reduction in resistivity due to an ordering induced by the stress
relaxation. Further annealing probably results in the flake
microstructure passing through metastable phases which have
resistivity values in the vicinity of those reported by other
investigators in Table 12. Additional experiments would need to be
performed to confirm this hypothesis.
Table 12. Resistivity Values cited in Literature
Bulk electrical Composition Temperature Reference Resistivity,
( C) ( ohm-cm )
8.50 x 10~2 In0.5Ga0.5Sb 30 Coderre and Woolley [92].
3.16 x 10 2 la-. CGSQ cSb 30 Omskii and Kblaniets [95].
1.99 x 10~2 InQ>5GaQ 5Sb 30 Omskii and Kblaniets [93].
7.90 x 10~2 Iitq 53Gao 47Sb 30 Woolley and Gillett[94],
1.10 x 10"1 In0.52Ga0.48Sb 30 Joullie et al.[78].
CHAPTER 5 /
SUMMARY AND CONCLUSIONS
The results obtained during the present investigation allow the
following conclusions:
1. Rapid solidification of several InxGa1_xSb compositions resulted in
the production of flakes or powders depending on the rpm of the copper
disk. Surface examination of the flakes showed presence of second phase
precipitates of the order of 1 micron which were rich in InSb. The
flake thicknesses ranged from 15 microns to 50 microns for varying disk
speeds. The powders obtained were irregular and varied in sizes from
one micron to several hundred microns.
2. Furnace cooled ingots yield a mixture of non equilibrium phases
which are retained in the powders and flakes on subsequent rapid
solidification.
3. X-ray and TEM analysis showed the flakes and powders to be
microcrystalline. The grain sizes varied With the disk velocity
exhibiting a minimum grain size of 2.7 microns in the case of the
flakes produced by rapid solidification in the intermediate velocity
range of 45-60 m/s. The surface of the flakes showed fine cellular
morphology whereas the cross sections of both the powder and the flake
showed presence of fine columnar grains. Powders obtained from rapid
solidification at high rpm showed presence of an amorphous phase which
97
98
is highly electron beam sensitive.
4. Thermal analysis indicated suppression of the metastable phase
showed by X-ray analysis in the case of the flakes and powders obtained
by rapid solidification in the intermediate velocity range of 45-60
m/s.
5. The flakes do not show presence of any substantial amounts of oxygen
or copper on the surface.
6. Bulk electrical resistivity measurements on flakes of composition
In0.51Ga0.49sb showed a resistivity minimum at an annealing temperature
of 100°C.
REFERENCES
1. Lendvay, E. Progress in crystal growth and characterization, Vol. 8, p. 371, 1984.
2. Ananthraman, T. R., and C. Suryanarayana. J. Mat. Sc., Vol.6, p. 1111, 1971.
3. Falkenhagen, G., and W. Hoffman. Z. Metallk. Vol.43, p. 69, 1952.
4. Duwez, P., R. H. Willens, and W. Klement. J. App. Phys. Vol.31, p. 1136, 1960.
5. Ananthraman, T. R. Metallic glasses. Production, properties and application. T. R. Ananthraman, ed., Switzerland, Rockport, Ma: TransTech publications, p. 1, 1984.
6. Guntherodt, H. J. Proceedings of the Fifth International Conference on Rapidly Quenched Metals. S. Steeb and H. Warlimont, eds. Elsevier Science Publishers B. V., p. 591, 1985.
7. Flinn, J. E. Rapid Solidification Technology for reduced consumption of strategic materials. Park Ridge, N. J., U.S.A.-Noyes Publications, p. 43, 1985.
8. Cohen, M., Kear, B. H, Mehrabian, R. Proceedings of the Second International Conference on Rapid Solidification Processing ± Principles and Technologies II, R. Mehrabian, B. H. Kear and M. Cohen, eds., Claitor's Publishing Division, Baton Rouge, LA, p. 1, 1980.
9. Jones, H. Proceedings of the First International Conference on Rapid Solidification Processing: Principles and Technologies, R. Mehrabian, B. H. Kear and M. Cohen, eds., Claitor's Publishing Division, Baton Rouge, LA, p. 28, 1977.
10. Mehrabian, R. Proceedings of the First International Conference on Rapid Solidification Processing: Principles and Technologies, R. Mehrabian, B. H. Kear and M. Cohen, eds., Claitor's Publishing Division, Baton Rouge, LA, p. 9, 1977.
99
100
11. Grant, N. J. J. Metal, Vol. 35, p. 20, 1983.
12. Jones, H. Rep. Prog. Phy., Vol. 36, p. 1425, 1973.
13. Predecki, P., A. W., Mullendore, and N. J. Grant. Trans. Met. Soc. AIME, Vol. 233, p. 1581, 1965.
14. Harbur, D. R., J. W. Anderson, and W. J. Maraman. Trans. Met. Soc. AIME, Vol.245, p. 1055, 1969.
15. Miroschnichenko, I. S. and V. 0. Zakharov. Industr. Lab., Vol. 35, p. 362, 1969.
16. Lohberg, K and H. Z. Muller. Metallkunde, Vol. 60, p. 231, 1969.
17. Lohberg, K and H. Muller. Fizika 2, Suppl. 2, paper 4, 1970.
18. Brower, W. E. Jr., R. Strachan, and M. C. Flemings. Cast Metals Res. J., Vol. 6, p. 176, 1970.
19. Ruhl, R. C. Mat. Sc. & Eng., Vol. 1, p. 313, 1967.
20. Jones, H. Mat. Sc. & Eng., Vol. 5, p. 297, 1969/1970.
21. Jones, H. Mat. Sc. & Eng., Vol. 5. p. 1, 1969.
22. Toda, T and R. Maddin., Trans. Met. Soc. AIME, Vol. 245, p. 1045, 1969.
23. Chen, H. S. and D. Turnbull. Acta Metall., Vol. 17, p. 1021, 1969.
24. Davies, H. A. and J. B. Hull, Scripta Metall., Vol. 6, p. 241, 1972.
25. Amorphous Glassy Metal and Allays, Tech. Rep. AFML-TR-78-70, 1977.
26. Beghl, G, R. Matera, and G. J. Piatti, Mat. Sc., Vol. 5, p. 820, 1970.
27. Pietrokowsky, P, Rev. Sc. Instrum., Vol. 34, p. 445, 1963.
28. Pond, R and R. Maddin. Trans. Met. Soc. AIME, Vol. 245, p. 2475, 1969.
29. Chen, H. S. and T. T. Wang. J. Appl. Phys., Vol. 41, p. 5338, 1970.
30. Chen, H. S. and C. E. Miller, Rev. Sc. Instrum, Vol. 41, p. 1237, 1970.
101
31. Chen, H. S., Leamy, H. J. and Barmatz, M. J. Non-Crystalline Solids, Vol. 5, p. 444, 1971.
32. Chen, H. S. Mat. Sc. & Eng., Vol. 25, p. 59, 1976.
33. Takayama, S J. Mat. Sc., Vol. 11, p. 164, 1976.
34. Pond, R. B. and J. M. Winter. Mat. Sc. & Eng., Vol. 23, p. 87, 1975.
35. Jones, H. Treatise on Materials Science and Technology, Ultra rapid quenching of Liguid Alloys, H. Herman, ed., N. Y., London, Academic Press, p. 18, 1981.
36. Bunge, H. J. Z. Metallkunde, Vol. 67, p. 720, 1976.
37. Kavesh, S. Rapid Solidification Technology, Source book, R. L. Ashbrook ed., Metals park, Ohio, American Society for Metals, p. 73, 1983.
38. Chen, H. S. and C. E. Miller. Mat. Res. Bulletin, Vol. 11, p. 49, 1976.
39. Maringer, R. E. and C. E. Mobley. J. Vac. Sc. Technology, Vol. 11, p. 1067, 1974.
40. Maringer, R. E. Rapid Solidification Technology, Source book, R. L. Ashbrook ed., Metals park, Ohio, American Society for Metals, p. 121, 1983.
41.Glickstein, M. R., R. J. Patterson, II, and N. E. Shockey. Proceedings of the First International Conference on Rapid Solidification Processing: Principles and Technologies, R. Mehrabian, B. H. Kear and M. Chen, eds., Claitor's Publishing Division, Baton Rouge, LA, p. 46, 1978.
42. Berkowitz, A. E., and J. L. Walter. Mat. Sc. & Eng., Vol. 55, p. 275, 1982.
43. Heinemann, A. Proceedings of the Fifth International Conference on Rapidly Quenched Metals. S. Steeb and H. Warlimont, eds., Elsevier Science Publishers B. V., p. 27, 1985.
44. Willardson, R. K. Compound Semiconductors - Preparation Of III-V Compounds, Vol. 1, R. K. Willardson and H. L. Goering, eds., Reinhold Publishing Corporation, N. Y., p. 8, 1962.
102
45. Hilsum, C and Rose-Innes, A. C. Semiconducting III-V Compounds. Vol. 1, Heinz K. Henisch ed., Pergamon Press, N. Y., p. 5, 1961.
46. Muller, E. C. and J. C. Richards. J. App. Phys., Vol. 35, p. 1233, 1964.
47. Stringfellow, B. Advances in Electronic Materials, B. W. Wessels and G. Y. Chin eds., American Society for Metals, Metals Park, Ohio, p. 11, 1986.
48. Breberick, R. F. Trans. Met. Soc. AIME, Vol.2, p. 1657, 1971.
49. Blom, G. M. and T. S. Plaskett. J. Electrochemical Society, Solid State Science, Vol. 118, p. 1831, 1971.
50. Woolley, J. C. and Smith, B. A. Proc. Phys. Soc., Vol.72, p. 214, 1958.
51. Neuberger, M. Handbook of Electronic Materials, Plenum Press, N.Y., Vol. 1-7, p. 17, 1972.
52. Antypas, G. A. J. Crystal Growth, Vol. 16, p. 181, 1972.
53. Rode, J. R., E. R. Gertner, A. M. Andrews, D. T. Cheung, W. E. Tennant.J. Electron. Mater., Vol. 7, p. 334, 1978.
54. Miki, H., K. Segawa, M. Otsubo, K. Shirahata and K. Fujibayashi. GaAs and Related Compounds, J. Blok, ed., Inst. Phys. Conf. Ser., London, Vol. 24, p. 16, 1975.
55. Ohta, K., M. Kawasuma, T. Tanuma and S. Kataoka. Jpn. J. Appl. Phys., Vol. 16, p. 1605, 1977.
56. Joullie, A., P. Esquirol and G. Bougnot. Mater. Res. Bull., Vol.9, p. 241, 1974.
57. Stringfellow, G. B. J. Crystal Growth, Vol. 27, p. 21, 1974.
58. Phillips, J. C. and J. A. Vechten. Phys. Rev., Letters, Vol. 22, p. 705, 1969.
59. Van Vechten, J. A. Phys. Rev., Vol. 182, p. 891, 1969.
60. Koster, W and B. Thoma. Z. Metallkunde, Vol. 46, p. 293, 1955.
61. Goryunova, N. A. and N. N. Fedorova.J. Tech. Phys. (U.S.S.R), Vol. 25, p. 1339, 1955.
103
62. Woolley, J. C. and D. G. Lees. J. Less Common Metals, Vol.1, p. 192, 1959.
63. Boltaks, B. I. Diffusion in Semiconductors, Translated by J. I. Carasso, H. J. Goldsmid, ed., N. Y., Academic Press, p. 306, 1963.
64. Ivanov-Omskii, V. I. and B. T. Kolomiets. Soviet Physics (Solid State), Vol. 1, p. 834, 1959.
65. McGroddy, J. C. and M. R. Lorenz and T. S. Plaskett.Solid State Commune, Vol. 7, p. 901, 1969.
66. Blakemore, J. S. Canadian J. Phys., Vol. 35, p. 91, 1957.
67. Plaskett, T. S. and J. F. Woods. J. Crystal Growth, Vol. 11, p. 341, 1971.
68. Hamaker, R. W. and W. B. White. J. Electrochemical Society, Vol. 116, p. 478, 1969.
69. Pfann, W. G. Trans. Met. Soc. AIME, Vol. 203, p.961, 1955.
70. Joullie, A., R. Aulombard and G. Bougnot. J. Crystal Growth, Vol. 24/25, p. 267, 1974.
71. Michel, J., R. Esquirol, A. Jouillie and E. Groubert.GaAs Related Compounds, J. Blok, ed.,Inst. Phys. Conference Series, London, Vol. 24, p. 94, 1975.
72. Miller, J. F. and R. D. Baxter. J. Electron Mater., Vol. 2, p. 4, 1973.
73. Rodot, H., A. Hruby, and M. Schneider. J. Cystal Growth, Vol. 3, 4, p. 305, 1968.
74. Yip, V. F. S. and W. R. Wilcox. Mat. Res. Bull., Vol. 11, p. 895, 1976.
75. Yip, V. F. S., C. E. Chang and W. R. Wilcox. J. Crystal Growth, Vol. 29, p. 69, 1975.
76. Wilcox, W. R., J. F. Yee, Mu-Ching Lin, K. Sarma and S. Sen.Sci. Technol., Ser., 38 (Sky Lab. Sci. Exp. Proc. Syrup.),, p. 27, 1974.
77. Sen, S. and W. R. Wilcox. Mat. Res. Bulletin, Vol. 13, p. 293, 1978.
104
78. Joullie, A., J. Allegre and G. Bougnot. Mat. Res. Bulletin, Vol. 7, p. 1101, 1972.
79. Yee, J. F., Mu-Ching Lin, K. Sarma and W. R. Wilcox. J. Crystal Growth, Vol. 30, p. 185, 1975.
80. Lefever, R. A., W. R. Wilcox, K. R. Sarma and C. E. Chang. Mat. Res. Bulletin, Vol. 13, p. 1181, 1978.
81. Woolley, J. C. and B. A. Smith. Proc. Phys. Soc., Vol. 72, p. 214, 1958.
82. Bachmann, K. J., F. A. Thiel, H. Schreiber, Jr. and J. J. Rubin. J. Electron. Mater., Vol. 9, p. 2, 1980.
83. Greene, J. E., C. E. Wickersham, and J. L. Zilko. J. Appl. Phys., Vol. 47, p. 2289, 1978.
84. Yano, M., T. Takase and M. Kimata. Jpn. J. Appl. Phys., Vol. 18, p. 387, 1979.
85. Gotoh, H., T. Yamamoto and M. Kimata. Jpn. J. Appl. Phys. Vol. 21, p. L7.67, 1982.
86. Syamalamba, Y., B. S. Naidu and P. J. Reddy. Proc. Nucl. Phys. Solid State Phys. Symp. Vol. 24(C), p. 515, 1981.
87. Syamalamba, Y., B. S. Naidu and P. J. Reddy. Phys. Stat. Sol., Vol. 60, K9, 1980.
88. Amemiya, Y., H. Terao and Y. Sakai. J. Appl. Phys., Vol. 44, p. 1625, 1973.
89. Wickersham, C. E., G. Bajor and J. E. Greene. Solid State Communications, Vol. 27, p. 17, 1978.
90. Weisberg, L. R. The Art and Science of Crystal Growing, J. J. Gilman, ed., John Wiley Publications, p. 380, 1963.
91. Underwood, E. E. et. al. Ceramic Microstructures and their Analysis, Significance and Production, R. M. Fulrath and J. A. Pask, eds., 1968.
92. Coderre, W. M. and J. C. Woolley. Canadian J. Phys., Vol. 47, p. 2553, 1969.
105
93. Ivanov-Omskii, V. I. and B. T. Kolomiets. Soviet Physics, Solid State, Vol. 1, p. 512, 1959.
94. Woolley, J. C. and C. M. Gillet. J. Phy. Chem. Solids, Vol. 17, p. 34, 1960.
95. Ivanov-Omskii, V. I. and B. T. Kolomiets, Soviet Phys., Vol. 2, p. 363, 1960.
96. Duwez, P., R. H. Willens and R. C. Crewdson. J. Appl. Phys., Vol. 36, p. 2267, 1965.
97. Laboratory manual for Integrated Circuit Technology,J. N. Fordemwalt, ed., Microelectronics Laboratory, Electical Engineering Department, University Of Arizona, 1985.
98. Carslaw, H. S. and J. C. Jaeger. Conduction of Heat in Solids, Second Edition, Oxford University Press (Clarendon), London, p. 286, 1959.