g
- z p 9!: / -;. f7}?
" 1 '1 [v r
[IHESIS g 52'»?- I l ”72.41%. fl
9' I a-A- .
?‘?‘~: ' 6942414
5 : sad; 5: , , '2; 3:: 1.931532
I
a. 3'2 . .;, V, l
n . l ' ’I ,l
:f "b’f’:? 9:423:01:
. 'J
%
vo~ _ ‘i
Vauzmmhmméw

This is to certify that the

thesis entitled

PHASE IDENTIFICATION AND
CHARACTERISATION 0F TITANIUM-62H

presented by l

Krishnan Narasimhan l

has been accepted towards fulfillment
of the requirements for

MaSterS degree in Meta] 1 I‘m—y l

 

 

 

 

Date—gglgg~

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

l
., i, 77 , .“x% - __.

 

 

 

MSU

LIBRARIES
—;—.

 

 

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

 

 

 

 

 

 

PHASE IDENTIFICATION
AND
CHARACTERISATION OF TI-6211

Krishnan Narasimhan

A THESIS

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

MASTER OF SCIENCE

Department of Metallurgy, Mechanics
and Materials Science

1984

ABSTRACT

PHASE IDENTIFICATION AND CHARACTERISATION IN
A Ti—6WTZAl—2WT%Nb—1WT%T3—IWTZMO ALLOY:

MORPHOLOGY, CRYSTAL STRUCTURE AND SOLUTE PARTITIONING
By

Krishnan Narasimhan

The purpose of this investigation was to determine the nature,
morphology and crystallography of various phases present in Ti 6211
as a consequence of various isothermal holding temperatures in the
alpha + beta range. Solute partitioning studies were conducted using
Scanning Transmission Electron Microscope and energy dispersive analysis
to determine the nature of elements stabilising the various phases
detected at a particular isothermal holding temperature. The phenomenon
of ordering in Ti 6211 was investigated and evidence has been presen—
ted regarding the formation of an ordered phase after aging at 450°C
for seventy two hours. TEM techniques have been used to establish the
crystal structure of various phases present at a particular iso—

thermal holding temperature.

ACKNOWLEDGEMENTS

I owe a deep sense of gratitude to Dr. Kali Mukherjee
for his guidance and encouragement.

I'm particularly indebted to Subasish Sircar. Narendra
Dahotre, Shekhar Subramaniam and Tong C. Lee for their
assistance in my project endeavours.

I wish to thank my other colleagues for their valuable
assistance when I needed it. and finally I'm indebted to the

office of Naval Research for funding the project.

a

INTRO

REVIE
2.1.1
2.1.2
2.1.3
2.1.#
2.1.5
2.1.6
2.1.7

REVIE
3.1

EXPER
h.1
h.2

TABLE OF CONTENTS

DUCTION ......................................
w OF PREVIOUS STUDIES.. ......................
Martensites in Titanium Alloy..............
Martensite Crystallography.................
Omega Phase......... ....... . ............ ...
Alpha Phase Formation ................ ......
P' Formation. .......... . ..... . ........... ..
0(2 Formation........ ......... . .........
The Interface Phase........................
W OF RELEVANT ELECTRON-ANALYTICAL TECHNIQUES.

Scanning Transmission Electron Microsc0py....
3.1.1 Microdiffraction in STEM..............
3.1.2 Electron Beam Microanalysis in STEM...
3.1.3 Principles of EDS System Operation....

3.1.4 Qualitative EDS Microanalysis.........

IMENTAL PROCEDURE... ........................
Heat Treatment...............................
Study of Heat Treated Samples................
4.2.1 X-ray Diffraction.....................

4.2.2 Scanning Electron Microscopy..........

Page

1

11
16
19
21
23
3h

42
1+2
1+3
#6
u7
48

53
53
53
53
53

TABLE OF CONTENTS -- continued

Page

5 RESULTS AND DISCUSSION............................ 56
5.1 X-ray Diffraction Studies......... ..... ...... 56

5.2 Transmission Electron Microscopy............. 59

5.3 Scanning Transmission Electron Microscopy.... 67
5.3.1 Analysis of Ordering in Ti-6211....... 67

5.u ‘Summary and Conclusions...................... 90

RLfiF'ERENCESOOOOOOO......OOOOOOOO0.000000000000000000000 92

’11
C)

I?)

U'\ (rm N

LIST OF FIGURES

Page
Partial phase diagram of beta Isomorphous
system.... ...... ......................... ..... .... h
Partial phase diagram of beta eutectoid system.... a
Phase diagram of Titanium-aluminium system........ 5
Titanium-tantallum system. ....... . ...... . ..... .... 6

Constitutional diagram of titanium tantallum
system............................................ 6
Constitutional diagram of titanium tantallum
system............................................ 7
Burgers mechanism for martensitic transformation.. 12
Lattice correspondence of martensites............. 13
SEM micrograph showing alpha...................... 14
Titanium-Alumnium phase diagram................... 25
Titanium-Alumnium Moly phasediagram............... 27
Provisional Ti-Al phase diagram................... 28
Do19 super lattice................................ 30
D30 results from Ti-15 at % Al.......... ...... .... 31
Ti-Al-Moo.25 isopleth............................. 33
Solute concentration profile in four regions...... 38
Schematic diagram of STEM.............. ..... ...... nu
Detector efficiency for 3mm silicon. .......... _.... #9
Escape depth for secondary electron and

X-raYBMiSSion..o. ..... 00000000000 ..... o oooooo 0000 51

20
21
22
23
2n
25
26
27
28
29
3O
31
32
33
34
35

36.

37
38
39
40

M
42
43
Lin

Heat treatment schedule.... ............. ............
X-ray diffraction peaks.... .........................
X-ray diffraction peaks.......... ................. ..
Schematic phase diagram of Ti 6211... ...... . ...... ..

Bright field and SAD for 950°C sample...............
TEM micrograph of 850°C sample......... .......... ...
TEM micrograph and SAD of 750°C sample.... ........ ..
Bright field and SAD for 700°C sample...............
TEM micrograph of 850°C sample......... ........... ..
SEM micrograph of 550°C sample........... ....... ....
SEM micrograph of 650°C sample......................
SEM micrograph of 850°C sample.. ........ ............
SEM micrograph of 850°C sample......................
SEM micrograph of 700°C sample......................
SEM micrograph of annealed sample....... ...... ......
SEM micrograph showing voids........................
STEM micrograph of 750°C sample.....................
STEM micrograph showing acicular martensite.........
STEM micrograph of 750°C sample showing twins.......
STEM micrograph of 750°C sample showing twins.......
Energy dispersive analysis (EDA) of 750°C

specimen........................ ....... ...... ..... ..
STEM micrograph of 650°C sample.....................
EDA of 650°C sample.................................
STEM micrograph of 800°C sample.....

STEM micrograph of 800°C sample.....................

Page
54
57
58
6O
62
63
64
65
66
68
68
69
69
7O
7O
71
71
72
72
74

73
74
76

77
77

1+7
48
49
50
51
52
53
54
55
56
57

EDA of 800°C sample................... ....... .......
EDA of 800°C sample from interface......
STEM micrograph showing 700°C sample................
STEM micrograph showing 700°C sample interface......

EDA of 700°C specimen.............. .................

EDA of 700°C specimen at interface......... ....... ..

STEM micrograph showing alpha platelet..

Magnified image of alpha2 precipitates... ..... ....

EDA from precipitate region............

EDA from region around alpha2 ................... ....

SAD of ordered alphaz. ............ ....

Page
79
80
82
82
83
84
86
86
87
88
89

 

1. INTRODUCTION

The use of Titanium alloys in structural applications is
undergoing continuous expansion and such an expansion results
in the requirement for alloys which have wider ranges of
properties and whose properties can be tailored to fulfill
the requirements of a specific application. Titanium 6211
(Ti-6Al-2Nb-1Ta-1Mo) is such an alloy which has recently
gained attention for many applications. This is primarily
an alpha alloy with some amount of beta. A comprehensive
research program was initiated to develop a better under-
standing of the effects of composition and microstructure on
the structure related properties of titanium alloys with
emphasis on Ti-6211. Though the knowledge of the physical
metallurgy and fracture mechanics of titanium alloys has
advanced considerably during the last four decades. several
technical issues with regard to this alloy have remained
unclear. The extent of interest in the microstructure of
Ti-6211 has been prOpelled by the fact that the mechanical
properties like strength. ductility, and fracture toughness
are strongly influenced by the microstructure. It is
believed that fracture toughness can be improved by raising
the solution temperature to produce a microstructure with
coarse widmanstatten alpha transformed from beta. In
addition to good mechanical properties, Ti-6211 has good

resistance to corrosion. A host of microstructural features

1

2

can be produced depending upon the hot working conditions.
heat treatment. and variations in quenching and cooling rates.
In such a two pahase alloy the martensitic transformation is
influenced by the chemistry of beta phase from which the
martensite is derived. The beta phase (b.c.c.) of titanium
alloys can transform martensitically into several crystallo-
graphic forms:

1) c' (h.c.p.) martensite

2) a" (orthorhombic) martensite

3) face centered orthorhombic martensite

4) face centered cubic martensite (observed only in thin

films [1]

The stability of the B-phase and the crystallography of the
martensite are related to the solute enrichment of the beta
phase. Whereas the martensitic transformation in a single
phase austenitejswith an unique habit plane and crystal
structure. the martensitic transformation in an alpha + beta
alloy can manifest variations in crystallography and trans-
formation kinetics depending on the solute partitioning
between the two phases. Stress induced martensite can occur
[2] if a metastable beta phase exists in the alloy as a
consequence of rapid cooling or a supersaturation of a beta
phase with beta stabilising elements at the solution treatment
temperature. The present investigation is concerned with the
detection of various phases present as a function of iso-
thermal heat treatment at various temperatures within the

alpha # beta phase field. Electron beam microprobe. and

3
energy dispersive microanalysis, by using high resolution
STEM have been used to determine the relative ratios of alpha
and beta stabilisers within the phases obtained after a given
heat treatment. Phase identification. morphology. and solute

partitioning in this alloy are discussed in this thesis.

 

 

 

Mahatma-u
“ohm
\
\ \ \
\
g \\ \ \
i... \ k ’3
e 0'5 \
.: \ K
I \ \
M. \M.
v t
l L_
memo-n

Fig - 1 - Partial Phase Diagram of the Beta

  
 

Isomorphous System .

Alloying elements of the beta momor-
phous type are V, Mo, Ta, and Ch.

5 . mom-um: Phen-

 

 

 

 

l
5 a o Internet-m: Phase
Allov Content
F1 0 . 2 . Partial Phase Diagram of the Beta

Eutectold System .

Alloying elements of the beta eutectoid
type are Mn, Fe, Cr, Co, Nl, Cu, and
SI.

 

In"! It! can ALquwuu
30 I :0

 

 

 

 

 

 

 

 

 

 

 

 

I C O I 0 IS 20 0 ‘0 70 IO 90
“00 L l 1 l l__ I l i I l lL-l I I
"20' 900
~:‘\
‘\‘ \ = -
‘ \::‘ ~ .: ‘00 F" "‘
Isoo \ ~\ i —
\
‘4‘
{\ 7w— 0.1, . —I
\
Isoo 2‘ “‘3. __ “s _
'< s:
e.\’ rut
'3, N «o ’ 1
“0° . \V ‘ O th‘kl’lz
. ‘\ \\
B / ‘\ ;.\3.°e
IIoo \

°c
4i
\

//L

 

 

 

R
\\

\

\

 

 

 

8
8
\
---1 .-..J tWI‘ifluJ‘--- I-
"1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

900
“J ‘
‘ 33.5 «.5
“9° IzuI ‘ 33.3) r
I I I n
m : I : 'I
I
I I ' h “V “o-
l | I g
I | I
too : I 5
I
I : I I ‘1“.
I I l I.
500 l I II
0 I0 20 so ‘0 so so to so so I00
II Home "new ALIIIIIIIIIII II.

Fig.3. Constitutional diagram of titanium-aluminum system.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Writ at“.
own” 00 so to N N no [005.00 b
1-) .IW0 1 _ I I I I I / .
z’ 4
J
I
I
>- I/ 4
r
l
. I
’ 4'7 /
r/
-..—AA [III / 4
‘VW .
"W
P l I
/
A u
" /
/
r
in; I, /
r’ _‘
" I .ma 3
I 3
‘ p. ’l 4 E
S
: 1/4 E
2 me .
S #494 c
Q F «1000
15;;
.- ‘IVW
I
' . um
1“ m! « mo
an 1
0 I0 30 JO ’0 30 . .30 70 IO 90 I00
WI“ eUJe
Fig.4.
x: . . f1
9 \ . '0. e 1
\ '\. ' , ' . _ \
I ‘\\. II
I \.
”a ' a.“
. \\.\\“
.. I 'N,\
i \,
--- , \
I \
° \
I l \
1 1 l
W, D 10 )0 .0
Win $114
Fig.5.

Constitutional diagram of titanium-
tantalum system.

0 .
IcIIIpcI .IIUI‘c L

Weight Percent Tentcicm "
30 13 20 10 £0 30 EC 7‘: 50 SO

I V 1' V 1' _ v v f

I
l)

 

290

 

—_

 

c '3 2': .io I; - :: .‘. ‘0 2::

TI Atomic ?er:::~.: 'cn'.:!;.-n Tq

000_
U
C

0

Fig.6. Ti-Ta Phase Diagram

2. REVIEW OF PREVIOUS STUDIES

Titanium alloys are broadly classified into alpha
titanium. beta titanium and alpha + beta titanium. depending
on the alloying elements present in the alloy. Elements like
Al. Ge. Ga. C. 0 and N stabilise alpha:Le.in the region where
alpha and beta coexist. these elements are more soluble in
alpha than in beta. Beta titanium alloys have predominantly
beta stabilising elements like V, Mo. Ta. Nb etc which are
completely miscible with the beta phase. or elements like Mn.
Fe. Cr. Co. Ni, Cu, Si (beta eutectoid elements) which
transform beta phase into alpha and an intermetallic phase.
Fig. 1 and 2 illustrate the above mentioned categories. Ti-
6211 has 6.0 w/o aluminum. 2.0 w/o niobium. 1.0 w/o tantalium
and 0.8 w/o molybdenum. Aluminum is an 'a' stabiliser and
the rest are beta stabilisers. Figs 3-6 illustrate the phase
diagrams of these elements when alloyed with titanium. Two
other elements zirconium and tin have the same extent of
solubility in both alpha and beta and are essentially

strengthening agents.

2.1 Phase Transformations in Titanium Alloys

Phase transfdrmations occuring in titanium alloys can be
classified into eight groups after Murakami [3]
1- Decomposition of beta phase when quenched to produce

alpha prime. alpha double prime, athermal omega, retained

beta.

2. Metastable beta transforming isothermally to isothemal
omega. alpha double prime. beta prime. alpha. a2, 82.

3. Decomposition of retained beta leading to isothemal
omega. beta prime. widmanstatten alpha. raft alpha. a2,
82, stress induced martensite. isothermal martensite.

4. Decomposition of martensite producing alpha. intermetallic
compound. and beta.

. Formation of alpha phase from omega. and from beta.

. Precipitation from alpha producing ordered phase 'a '.

5

6. Eutectoid decomposition.

7 2
8

. Formation of interface phase.

2.1.1 Martensites in Titanium Alloy

High temperature b.c.c. phase termed 'p' transforms
martensitically during quenching. which can be classified as
hexagonal martensite [4. 5]. face centered cubic martensite.
orthorhombic martensite [6] and face centered orthorhombic
martensite [7]. a' martensite (hexagonal) is the most
prevalent martensite. Two morphologies of the same have been
observed 8 massive martensite and acicular martensite [9,]
flxfl. The massive martensite has colonies which can be resolved
optically,exihibits surface relief effects from plates within
the colonies. The boundaries between the plates have been
shown to consist of walls of dislocations, whereas the
interior of the plates contain a relatively dense tangle of
dislocations. Massive martensites have also been called

packet or lath martensites. With increase in solute con-

centration,the massive martensite colony size decreases. and

10

at a Specific solute concentration (which depends on the
alloy systems) the colonies become individual plates. These
plates have an acicular morphology. This martensite is
frequently internally twinned on {10I1}a-. Habit plane
determinations on the acicular martensite have been performed
and these martensites have. for the most part, the {334}B
habit plane [11. 12] which is predicted by the Bowles and
McKenzie's crystallographic theory of martensite.

The second type of titanium martensite (a") has an
orthorhombic structure. This structure was first discovered
by Bagariatskii in 1959 who reported it on the basis of X-ray
diffraction. This martensite occurs in some binary alloys.
eg Ti-Mo, Ti-Nb. Ti-W. Ti-Re. and has been also observed in
Ti-6Al-ZSn-4Zr-6Mo accompanied by a reduction in ductility
[13]. The lattice parameters of orthorhombic martensite are
strongly composition dependent. Characteristic pattern
reveals five low angle lines compared to three for hexagonal
martensite. Orthorhombic martensite has been examined in the
electron microscope in a Ti-8.5 Mo-0.5 Si alloy and has been
found to be internally twinned on {111}O planes [14].

Several workers have observed a face centred cubic
martensite in Ti-Fe [15], Ti-Cr [16] etc. This martensite
has always been reported on the basis of thin foil electron
microscopy evidence. Recent evidence obtained by Williams
and Rhodes showed that this martensite is orthorhombic in its

lullk form [17]. Strong evidence as a consequence of work

dcnae by several investigators indicates that the face centred

11

cubic and face centered tetragonal martensites reported on
the basis of thin foil electron microscopy, are thin foil
artifacts and are probably a thin foil version of a bulk
orthorhombic martensite.

Apart from athermal type of titanium martensites.
martensite forms during stressing in metastable bcc Ti alloys.
There is evidence of mechanical twinning [18. 19] in some
alloys, and it is suggested that the concentration and the type
of solute has an important influence on whether the stress-
induced product in a given alloy is twinning or martensite.
Electron microscopy of Ti-11.5Mo-4.58n-6Zr; Ti-14Mo-3Al [20,
21] showed bcc twins. There is enough evidence to show that
other alloys do form stress induced martensite and in these
cases the martensite product is orthorhombic. Difficulties
due to line broadening as a result of straining. such as poor
resolution. led to a study by Williams [21], who reported
that hexagonal stress induced martensite in different alloy
compositions could not be reproduced. but instead revealed
orthorhombic alpha double prime phase. It has been suggested
that alumium additions to Ti-Mo and Ti-V alloys promote
stress induced a". Some B-Ti alloys were reported to have
fcc stress induced martensite product 22 . and hcp

martensite [23].

2.1.2 Martensite Crystallography
Several workers have tried to explain the crystallography
<3f martensites a' and a" with respect to the parent phase

‘beta. Oka and Taniquichi [24] suggested that Mo content

 

12

.3: L3: .5388: E 33»: umxunauomopo $835.3. 3 33a: ozzu .5me
um..3=mo-.€on no coSsanmcnb no» «39:5 B 3333 52:23... mo 55953—2 .

.2...

      

a til «....2.
TxSII «EAI

           

«EAI «or...
LN-U chfi

...,4
9%.

I
V-

 

N
N

u
l

..-...-.—-.

\ «T‘Nn

\\\ .~-.u
:z.

 

:5 ,
«...: .0. as

a 3

14:;

13

X “0013

 

 

 

 

 

 

 

 

X Z [OOIIfi
’,,4r

 

\

YIOIOIfl

Lattice correspondence of martensites showing the four
unit cells of the BCC crystal and its face centered
tetragonal equivalent.

Fig.8.

14

\
\.

\;.

pairs 5.?”

Fig (9). SEM micrograph of 700° C specimen showing

 

widmanstatten alpha

15

influenced the lattice parameter of stress induced orthorhom-
bic phase to approach that of the beta phase. Sasano et al.
[25] confirmed that with increasing beta stabilising agent
the crystal structure of Ti martensite changes from th to
orthorhombic. Fig.7 illustrates the Burger's mechanism for
the beta (bcc)-—-alpha prime (hop) transformation in Ti-
alloys [26]. Mukherjee and Kato [27] suggested that all the
titanium martensites can be generated from a face centered
tetragonal equivalent of the four unit cells of the bcc beta
phase (Fig 8). In the transformation of beta -— alpha prime
Or alpha double prime. it was suggested [27] that for alpha
prime. the Bain strain is brought about by Burger's mechanism
where the Operation of the {112} <111>B shear systems
followed by a small volume increase results in the (0001)a'
plane from the (011)B plane. In addition to the above bain
strain.a shuffling on every other (011) plane is necessary
to create the hcp structure. In the formation of alpha
double prime martensite. after the operation of the {112}
<1111>B shars, the (011)B plane does not completely change
into the hexagonal close packed plane but stops midway
between the (011)B and (0°01)a' leaving the angles 61 and 62
different from 120°. Thus according to Mukherjee
and Kato the alpha double prime martensite can be considered
a distorted alpha prime martensite. In f.c.o. martensite
there is no shuffling and lattice correspondence is very

similar to Au-Cd martensite. From the Bain strains

calculated in Ti-12.6V by Mukherjee and Kato [27] for the

16

f.c.o. martensite it is obvious that f.c.o. martensite cannot
be considered a distorted alpha prime martensite. Recently
Sircar. Krishnan. Mukherjee [28] showed that the lattice
parameter of the beta-phase at the MS temperature is a
dominant factor which determines the stability of the beta-
phase and/or the type of martensitic transformation in Ti-
6211 alloy. The magnitudes and signs of the Bain-strain
associated with the martensitic transformations are directly
related to the beta-phase lattice parameter. From
theoretical analysis it was suggested that those beta stabi-
lising elements which reduce the unit cell size of the beta
phase. tend to favour the a" martensite. Mukherjee and Kato
[27] have also shown by strain energy minimization criteria
and by Bain strain calculation,that there exists a critical
Bain strain at which a transition from a' martensite to a"
martensite is favourable. This reveals the fact that the
transformation of beta -— alpha double prime is more
favourable than beta -— alpha single prime for certain
critical sclute concentration of the beta phase. 0f the
various alloying elements in Ti-6211 molybdenum produces a
contraction in the beta lattice parameter and thus tends to

favor the alpha double prime martensite formation.

2.1.3 Omgga Phase
Beta phase in titanium alloys when retained in meta-

stable form by quenching. can partially decompose during

quenching into another metastable precipitate known as the
athermal omega phase [29]. Such alloys which tend to

17

decompose to "w" phase contain a sufficient concentration of
beta stabilising elements to depress the martensite finish
temperature beneath room temperature. There is a very
limited composition range over which omega phase (athermal)
occurs. and the extent depends on the particular alloy system
under consideration. There is enough evidence to suggest
that athermal omega forms without a change in composition.
The structure of the omega phase is not cubic as suggested by
Frost et al. [30]. however it is less clear whether the phase
is hexagonal as claimed by Silcock [31] or trigonal symmetry
as claimed by Bagariatskii [32]. Athermal omega was
considered for sometime to form by a martensitic trans-
formation largely because its formation cannot be suppressed
by quenching. TEM studies have shown that athermal omega
occurs as extremely small particles ( 20-4OA°) with a very
high density. De Fontaine [33] proposed a model for the
formation of omega phase based on diSplacement controlled
mechanism. 'u' lattice can be generated from the beta phase
by a series of 2/3‘<111>Ilongitudinal displacements.

Further. Paton. de Fontaine and Williams [34] have used cold
stage transmission electron microscopy to show that the omega
phase can be reversibly formed on cooling beneath room
temperature. Such reversibility is consistent with a
displacement controlled transformation which can occur
without change in composition. This mechanism also explains
the complex diffuse intensity of diffraction observed-in

metastable beta alloys. The streaking in electron

18

diffraction pattern has been attributed to random or un-
correlated displacements of the atoms so that the omega phase
lattice is not generated but the periodicity of the bcc
lattice is reduced. Sass [35] has suggested that the diffuse
intensity results from coherent scattering from rows of
athermal omega phase particles resulting in sheets of
intensity lying normal to '<111). It has been reported that
variations in oxygen content of the alloy has a direct bearing
on the apparent composition limit at which sharp or well
defined omega phase reflection maximum can occur. leading to
a disagreement amongst various workers regarding the
composition at which athermal omega can be suppressed.
Another observation made is the case with which alpha
nucleates from unstable beta when there is a presence of
intermediate unstable products like ”m”.

In addition to athermal omega. isothermal omega can also
form as a metastable transition precipitate in alloys of
intermediate concentration of beta stabiliser during aging at
temperatures up to 550°C. This phase was originally
discovered in 1954 by Frost et al. [36]. Isothermal beta -—
omega transformation occurs most rapidly of all beta phase
decomposition modes, and the rapid kinetics of omega phase
formation are related to the case with which the hexagdnal
omega phase can be formed from the bee lattice. Isothermal
omega forms during aging as coherent [37] precipitates of
small size and high particle density [38. 39]. Ellipsoidal
[40] and cuboidal [41] omega phase morphologies have been

’I'Lfl" "I'- ha” 1" I

19
observed in titanium alloys. the morphology being controlled
by the misfit between the omega precipitate and the bcc
mattrix. Ellipsoidal omega phase occurs in low misfit alloys
like Ti-Mo. Ti—Ta. Ti-Nb. the long axes of the ellipsoids of
revolution lying parallel to a.<111>8 direction. Cuboidal
omega phase is observed in high misfit alloys such as Ti-Fe,
Ti-V, Ti-Cr, Ti-Ni. Ti-Mn. The cuboids form with their cube
faces parallel to {loolfgplanes [42]. Ternary additions like
Al. Zr. Sn influence the formation and stability of omega
phase. Studies have shown [43] that the principal influence
of aluminum on the stability of omega phase is to promote
separate nucleation and growth of alpha phase after
relatively short aging times: this alpha phase grows and
consumes the 8+!» mattrix. limiting the time temperature
range over which omega exists. Sn. or Zr additions reduce
the volume fraction of omega and stabilise the beta phase.
Such ternary additions reduce the maximum temperature of
stability of the omega phase and thus they are used to promote
omega - alpha transformation. Oxygen also influences omega
formation and stability by slowing down the kinetics of its
formation and reducing the volume fraction of omega phase
formed. Limited data suggests that oxygen tends to increase
the stability of the beta phase and does not significantly

promote the independent nucleation of alpha phase.

2.1.4 Alpha Phase Formation
Equilibrium alpha phase in titanium alloys can be formed

from the metastable beta phase, omega phase. B'-phase. or

20

during the decomposition of titanium martensites on tempering.
Morphology and distribution of alpha phase depends on the
alloy content and in the manner in which it forms. 0n
continued aging of alloys which have undergone the B t 8'
phase separation reaction. alpha phase can be nucleated
within the 8' regions and can either grow as ”rafts" of
random shape which contain a large number of very small
particles, or as well defined needles which grow with their
long dimension parallel to {110}B plane [44]. Alpha phase
can be nucleated directly from the metastable beta phase in
concentrated alloys aged above the maximum temperature for 8'
formation. In such cases two types of alpha phases are
observed. these type being different in morphology and
crystallography [45]. The first type of alpha phase known as
Widmanstatten alpha forms as plates and needles, the long
axes of which lies parallel to {110}B. This type-of alpha
phase forms with the characteristic Burgers orientation
relation. with each Widmanstatten plate is a single variant
of the Burgers orientation relation. This type of alpha is
observed in binary alloys linear in beta stabiliser and in
alloys which contain a good amount of aluminum eg. Ti-6Al-
4V. Recently Sircar. Krishnan. Mukherjee [28] have shown the
occurence of prominent Widmanstatten alpha plates in Ti-6211
Fig (9). Typical basket weave alpha structure, and blocky
spngetti alpha have also been observed in this alloy [28].

The second type of alpha phase occurs frequently in

binary alloys with a high concentration of beta stabilising

21

elements eg: Ti-18 wt.% Mo. Ti-8Mo-8V-3Al [46]. This alpha
is known to form as aggregates of very small particles, and
two distinct morphologies of this type have been observed.
one assuming a long lenticular shape and the other raft and
patchy. Selectedauea.diffraction techniques have shown that
the orientation relation between the alpha phase and the
beta matrix is not the commonly observed Burger's relation-
ship. Alpha regions contain alpha particles from more than
one variant of the orientation relation. suggesting that
sympathetic nucleation [47] is involved in the process. In
alloys aged at temperatures which result in direct formation
of the alpha phase, the distribution of the alpha phase is
inhomogeneous and plate size is coarse. This results in
large sections of untransformed beta phase between the coarse
alpha platelets. leading to a low tensile ductility in this
alloy. Such a deletmdous distribution can be improved by
defamation prior to aging. because such a deformation intro-.
duces a high dislocation density from which alpha can
nucleate homogeneously. leading to enhancement of properties

in beta phase alloys.

2.1.5 3} Formation

In titanium alloys the bcc phase can separate into two
bcc phases of different composition. This composition reac-
tion was originally termed 81 t 82 by some workers. It is
proposed that the mixture of bcc phases resulting from the
phase separation be called (8 + 8'). The B'-phase occurs as

uniformly distributed. coherent bcc zones in the beta phase

22

mattrix. in the range of ZOO-500°C temperature. The
morphology of the 8' regions is variable and appears to
depend on the composition difference and misfit between the
two bcc phases [48]. The occurence of 8' has been observed
in a wide range of alloy systems where sufficient beta
stabilisers preclude omega phase formation during low
temperature aging. The phase separation is thus viewed as a
kinetically favorable mode of beta phase decomposition in
alloys where direct formation of the equilibrium alpha phase
is sluggish and omega phase cannot be formed. The detection
of B' precipitates has so far been possible by thin foil
electron microscopy only because X-ray diffraction led to
peak broadening. The composition of B' have not been
determined directly. However. on continued aging. the
equilibrium alpha phase precipitates within the 8' regions,
revealing the fact that 8' is solute lean. The thermo-
dynamics of the phase separation reaction have been considered
by several workers. in particular Kauffman and Bernstein [49]
who seem to agree that the reaction requires an inflected
beta phase free energy composition curve. and that B' is a
metastable constituent. The stability and kinetics of
formation of the B' precipitates depend on the alloy
composition. The addition of ternary elements which promote
alpha nucleation i.e. oxygen and aluminum reduce the time and
stability of the B' precipitates by accelerating the rate at
which alpha phase is nucleated. One interesting aspect of

this transformation is that the B + 8' reaction followed by

23

the B' -— alpha reaction appears to be a promising means of
producing a uniform diSpersion of alpha hardening precipitates

in beta alloys. especially for deep hardening applications.

2.1.6 32 Formation

 

In alloy systems which exhibit an increasing solid
solubility in the alpha phase with increasing temperature, it
is possible to produce a supersaturated solid solution which
will subsequently undergo alpha decomposition during aging.
Several eutectoid alloys exhibit extensive solid solubility
and in these the alpha phase can precipitate the intermetallic
compound on aging below the eutectoid temperature. In two of
these systems. Ti-Si and Ti-Cu. significant supersaturation
can be achieved by quenching from under the eutectoid
temperature. The other important class of alloys in which
significant supersaturation can be achieved are alloys between
Ti and group III or group IV elements. notably Al. Ga. In, Sn
and Pb. All of these alloys decompose during aging to form
an ordered phase designated cg; this phase is based on the
composition T13X and has a D019 structure. The mode of
formation of the o2 phase depends on the alloy system under
consideration. however in Ti-Al alloys, the phase forms as a
uniform dispersion of coherent precipitates. As the 02
particle size increases, either by coarsening in dilute
alloys or by aging at high temperatures in concentrated
alloys. the a2 precipitates assume an elongated ellipsoidal
morphology, the direction of elongation being [b001]a 50

presumably because the alpha/02 misfit is smaller in the T

24

direction. The aspect ratio of the precipitates can be
altered in Ti-Al alloys containing ternary additions of Ga
and Sn since these two alloying elements have the opposite
effect on the T and a direction misfits of the ca phase [51].
In Ti-Al-Sn alloys the c2 phase is nearly equiaxed and in
Ti-Al-Ga alloys the °2 phase grows out as extended rods.
Considerable dispute existed between workers regarding the
extent and shape of the o2 phase field on the Ti-Al phase
diagram. Crossleys diagram shows that the Ti3Al phase exists
only as a stoichiometric line compound. whereas Blackburn's
diagram shows that the alpha2 phase exists as a single phase
over a wide range of composition. It has been suggested by
Williams that the shape of the o2 phase in Blackburn's
diagram is correct because existence of an extensive a2 phase
field has been confirmed. Blackburn's diagram Fig (10)
indicates that there is a anomalous contrast represent-
ing the limiting aluminum concentration at which the single
phase a2 structure exists. Blackburn suggested that the
anomalous phase contrast may be due to excess Ti atoms in the
antiphase boundaries. or due to a very thin layer of
disordered material.

Increasing amounts of ternary alloying elements lead to
a reduction of the volume fraction of the ordered az-phase in
quenched alloys [54]. This effect is understood qualitatively
if it is clear that these elements all lower the MS
temperature of titanium alloys. In addition to an influence

on °2 volume fraction. alloys that contain 5% at Nb also

IBOO

1100

900

700

500

300

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O aphose
e a+oz2
o az-phose //
§ Region of Anomalous +3 /
APB Contrast /°‘ D
B _ / o
// 000° "‘
/ °
// g
o co . o
a
014-012 . 0‘2
oeooos
. . [I
I
’ SRO
5 IO I5 20 25
—-—~>-

Atomic % Aluminum

Fig.10.

30

26

exhibit recrystallisation reaction. The recrystallisation
reaction frequently is initiated at prior beta grain
boundaries. possibly because such sites are a good source of
the disordered boundary which surrounds the recrystallised
regions.

Recently Namboodri [55] suggested a shifting of the
a-(otaz) transition point to lower aluminum concentrations
and to higher temperatures than those proposed by Blackburn.
Regions of short range order were determined by electrical
resistivity measurements. Namboodri et al. [55] investigated
the decomposition behaviour of the alpha phase in the Ti-Al
system containing up to 19 at % A1 (11.8 wt % Al).
Isochronal and isothermal resistivity changes were measured
in alloys originally quenched from temperatures above the
o-(o+o2) transition point. Precipitation of the ordered 02
phase followed typical precipitation kinetics and resulted in
a decrease in the as quenched electrical resistivity of the
alloys. Short range ordering raised the resistivity of
quenched alloys. and the plastic deformation of short range
alloys (ordered) led to decreases in the electrical resisti-
vity. On the basis of these studies, Namboodhiri et a1
suggested that the cb(crw§) boundary should be shifted to
lower aluminum concentrations than those proposed by
Blackburn. Recently, R. D. Shull, et al. [56] have

identified the presence of C2 (ordered Ti Al type) in Ti-6211

3
alloy. Till this data the presence of ch in titanium 6211

has not been unequivocably established. The fact that C?

27

Ti

 
   
   
 
 
 

 

v\ {TI-62H
o.\° n+8
V 20
o+oz+fi
30 . “*3
eg+fl+fiz,

 

700°C

0
Fig . 11 . ‘rl-rich corner of the TI-Al-Ho phase diagram at 700 C.

TEMPERATURE ('CI

253

 

I800 '-

|400 '-

I000

600 "'

 

LIQUID

 

 

l

 

 

 

l

 

:1-90

IO 20 30

ATOMIC 96 AL

40 50

Figure l2. Provisional Ti-AI Phase Diagram (Titanium-rich end).

.29

plays a prominent role is illustrated in fig (11) where the
Ti-rich corner of the Ti-Al-Mo phase diagram at 700°C is
revealed. It is recognised that these phase relationahips
are affected by the unintentional addition of interstitial
impurity elements like oxygen. hydrogen. and nitrogen. R. D.
Shull. et a1 [56] reported experimental data on the binary
Ti-Al system which were useful in locating the q/o+o2 phase
boundary in this system fig (12) by filling in the gaps in
previously reported data. Conclusive evidence was provided
for the detection of the o z. °2 order/disorder transformation
using DSC (differential scanning calorimetry). Extreme
precautions were taken to eliminate as much as possible
hydrogen, oxygen. during the manufacture of Ti-6211 for
experimentation. The range of aluminum compositions studied
was between (10 at percent to 21 at percent). Selected area
diffraction was used to confirm superlattice ordered spots.
In the disordered hexagonal phase aluminum atoms are
substituted randomly for titanium atoms, and in the ordered
hexagonal (a2 phase) of Ti3Al-type (D019 structure). there
are definite positions for the aluminum atoms fig (13). The
a2 phase is coherent with parent lattice. the mismatch in
lattice parameters being a few percent. Because of the
lattice similarities. the electron diffraction patterns of
the two phases differ only in the apperence of additional
superlattice Spots when o2 is present. Fig (14) illustrates
the DSC data with H = 4.2 J/g highlighting a It a order

2
disorder transformation. R. D. Shull, et al. [56] also

i L,

 

 

 

 

 

x

ted with the

The unit cell associa

DO..- typc superlattice.

Fig.13.

31

080

fiAH ' 44.2 J/g

 

000
700
600'0

".

 

Heating and Cooling Roie

 

 
 
 

 

    

29.9°C/min.

AH-4za/g .. A’90“ atmosphere

AH-43.5Jlg

AH'45.6J/g

9 O

0
§ 8 g s g g

r.

Ti-IS "lo Al

 

AH- 44.7 Jlg

 

 

Fig. 14° Differential scanning calorimetry data obtained on a
Ti-lS at.: Al sample aged at 625°C for 378 hours and on an
as-received sample of alloy Ti-62ll.

32

conducted experiments on Ti-Al-Mo alloys containing a constant
a at. percent Mo and Al contents ranging from 10 to 21 at.
percent. It was found that for these ternary Ti-Al-Mo alloys
the same a2 precipitates appeared as in the case of binary
Ti-Al alloys. The size of the o2 precipitates ranged from a
few nanometers to several hundred nanometers; the higher the
temperature and the closer the heat treatment to the phase
boundary. the larger the precipitation size. The criterian
for identifying equilibrium o2 precipitates is the observation
of distinct superlattice Spots coupled with dark field
images that show homogeneous distribution of alphaz. Fig (15)
shows the results for the ternary alloy Ti-Al-Mo 25. The
filled circles indicate those samples where equilibrium o2
was found as stated above. and the open circles indicate
those samples where equilibrium o2 was not observed. The
thin background lines indicate the binary Ti-Al phase
diagram (Fig 12). The thick solid lines indicate the o/a+a2
and alpha/alpha+beta boundaries determined for this system.
It is suggested that the q/a+a2 boundary shifts towards
higher Al contents on the addition of i at. percent Mo.
Specifically the presence of Mo shifts the q/a+o2 boundary
toward higher Al contents and shifts the alpha/alpha+beta
boundary toward lower temperatures.

Recently. we observed o2 precipitates in Ti-6211 after
an aging at 450°C for 72 hrs confirming the presence of
ordering in this alloy. Details will be furnished in the

results and discussion section of this thesis.

33

 

"00--

900

 

 

 

 

 

a
I.
'7()()- <3 0 .4
- a2
" 1. 1| 0| 1. '8
I I I I
°°°o IO 20 so
n I was ATOMIC % AL

99.75

Fig.15.

Ti-Al-Hoo.25 is0pleth. Thin background lines indicate
the diagram for the binary Ti-Al system. Data points
from this study for the a+o2 (O). o(o). 0+8“) phase
fields are also shown.

34

2.1.7 The Interface Phase

As early as 1966 the nature and structure of the
interphase boundaries between the alpha and beta phases in
n+8 titanium alloys was recognised as complex, but no
detailed analysis was attempted. Since then more information
is available regarding the nature and crystallography of the
interface phase which nevertheless is controversial. A
profound_ interest in this region stems from the fact that
the interface phase plays a significant role in various
mechanical properties. The alpha/beta interfaces may
influence the fracture behaviour of Ti alloys since they can
provide crack paths for fatigue failures, and tensile
overload failures. Strength and ductility may be affected if
the presence of the interface phase inhibits the transfer of
slip between the alpha and beta phases. Rhodes and Williams
[57] have reported that in Ti-6Al-4V alloys the interfacial
region is seen to be composed of a densely striated layer.

It was at first suSpected to be titanium hydride but this
conclusion was not confirmed after further experimentation.
Titanium hydride shrinks and disappears when heated to a
temperature above the solvus which was not exhibited by the
interface phase. Rhodes and Williams [57] summerised the
fact that the interphase face that forms in alpha/beta
boundaries is neither titanium hydride nor simply a region of
high dislocation density, but it is a distinct layer which
develops either as hexagonal-alpha phase having a different

orientation than the primary alpha into which it grows. or as

35

an fcc phase depending upon heat treatment. The occurence of
the complex interphase boundary appears generally in all.
titanium alloys containing alpha/beta boundaries which have
formed at temperatures in the range 675 to 875°C. Margolin
et al. [58] have postulated that the interface phase is
related to the precipitation of alpha. and that the thickness
of the interface structure is approximately the same as the
thickness of Widmanstatten alpha and the alpha projections
which emanate from the coarse Widmanstatten alpha. Formation
of the interface phase generally appears to take place when
precipitation of alpha would be expected to occur.
Precipitation of alpha from beta involves volume expansion
since alpha has a lower density than beta [59]. the
difference in size most be accomodated by both phases and
hence volumetric strains have to be considered in the
formation of the interface phase. and may contribute to the
formation of the fcc phase noted by Rhodes and Williams. The
presence of a fcc structure, twinned alpha or a heavily
dislocated layer of alpha of different orientation than the
contiguous alpha suggests that these interface structures
need not produce embrittlement of alpha-beta alloys [58]
several observations support this point of view. Slip lines
have been observed to pass readily across this interface from
alpha to beta under both tensile and fatigue loading
conditions. a process which suggests that the interface
structure. presumably present, is not a significant barrier.

It was suggested by Margolin et al. [58] that the interface

b;

36

phase may contribute to void formation by promoting
incompatibility between alpha and beta. But by itself this
assumed enhancement in void formation potential would not
necessarily lead to decreased ductility. and the largest
degree of incompatibility would be promoted at interfaces of
equiaxed alpha and beta [58]. Rhodes and Paton [60] studied
the conditions for formation of the interface phase in Ti-6Al-
4V. Their previous work had indicated that the interfacial
layer could form either with an fcc structure or with an hcp
structure. and evidence was presented that the interface
phase was alpha phase in a "non-Burgers" orientation or in a
Burgers orientation different from that of the adjacent
primary alpha. They also concluded [60] that the interface
phase does not grow during isothermal treatments. but rather
grows only during cooling from elevated temperatures to about
650°C. The width of the interfacial layer is a function of
the cooling rate. having a maximum of about 4000°A at 28° C/h.
The interface phase forms initially with an fcc structure
which subsequently transforms to hcp alpha phase. The layer
has been shown previously to occur either with a dense
internal structure (called striated alyer) or with no
internal structure (called a monolithic layer). The striated
layer was interpreted as having the hop structure of the
alpha phase, but in a different orientation from the adjacent
primary alpha. The monolithic layer was shown to have an fcc
structure [60] and obeys the crystallographic relationship
(110)B // (111)fcc // (0002)a

37

and [111]B // [110]fco // [1120101
It was suggested that the foe monolithic layer phase may be a
transition structure in the B-cxtransformation [60], and also
that the interface phase may result from sluggish beta
stabiliser (vanadium) diffusion away from the transforming
(beta -— alpha) interface region. Williams studied the
occurence of interface phase in Titanium-6211. and fig (16)
reflects a concentration profile of solute in four regions of
beta, fcc interface phase. hop interface phase, and alpha
phase. The conc. range of each phase is illustrated by two
dashed lines. Darkened lines o1 and a2 define the
concentration range of both hop interface phase and alpha
phase. Since they have the same crystal structure, the
concentration discontinuity is not expected to exist between
them. but this suggestion by Williams is prone to speculation
because concentration difference could exist as a consequence
of elastic misfit energy which can give rise to difference in
chemical potential in two regions of a single phase. This
concentration profile according to Williams can be applied
to Specimens which were beta solution treated and cooled to
a temperature below 840°C (where the fcc interface is
stable). It can also be applied to beta worked material
(as received) which has been n+8 solution treated at
temperatures where fcc interface phase is stable. eg: 700°C.
Williams showed that in Ti-6211 the alpha/beta interface
structure consists of two different layers of interface phase.

a monolithic layer (fcc structure) and a hop layer with dense

38

 

 

 

 

 

 

 

 

 

 

 

 

 

l
The equilibrium solute cone. 0: B at RT.

ll 81
The solute cone. 0: B at the critical temp,
below which lP ( FCC) is 'stable.

B.
Solute
COHC. | IF,
I
Nb } ”=2
Mo I
Ta
l
i . a I C11
I I
I g I
l
5 l 1 C12
: FCC l HCP l .
I | ‘
B J< lnterlace >l< Interfacek C1
. 1 Phase : Phase :
l l l ,
Distance :

 

Fig,]j§,The schematic description of solute concentration profiles in
. four regions consisting of B, fcc interface phase, hcp interface
phase, and a phase.

39

internal structure. The fcc interface is unstable at 8h0°c
and above. and the critical temperature below which fcc
interface phase is stable is believed to be between 820°C and
840°C. The concentration of beta stabiliser elements (Nb.
Mo. Ta) in beta. fcc interface phase (IF). hcp interface
phase (IH). and alpha vary in the following order:
B>IF>IH>a

According to Williams the most often observed orientation
relationship of alpha. fcc interface phase. and beta phase is
as follows

(0002).. // ($00)ch //(oi1)B

[1150] a // [011] fcc // [111] B
Banerjee and Arunachalam 61 studied the interface phase in

titanium alloys like BT and IMI 685 and found the interface

9
phase tobe thermally unstable at 873 K. and the time taken for
complete dissolution of the phase as suggested by them is

64 hours at this temperature in ET The Rhodes Paton model

9.
[50] for the formation of the fcc phase has been shown to be
inapplicable in these alloys. and that the fcc interface
phase is titanium hydride. Banerjee and Arunachalam [61]
visualised the formation of titanium hydride as follows:

As the beta - alpha transformation proceeds on cooling the
alloys through the transus. partitioning of various alloying
elements occur. The partitioning of hydrogen must be

particularly strong below 573 K where the solubility of

hydrogen in alpha decreases rapidly. Hydrogen atoms

migrating from the alpha phase to the interface phase will

40
precipitate in this region since the supersaturation exceeds
the solubility limit. The precipitation event takes place
when the beta -— alpha transformation is complete. Recently
Banerjee. Williams et al. [62] have suggested that the inter-
face phase (face centred cubic or hexagonal structure) is an
artifact produced by the electrolytic technique used in thin
foil preperation. Results were based on a near alpha alloy
Ti-6Al-1.SZr-3Mo-0.253i and Ti-10V-2Fe—3Al (metastable beta
alloy). Foils from these samples were prepared by jet
polishing in a solution containing 6% H2301+ in mathanol at
223 K in a fischione jet polisher or by ion milling from the
bulk. Different starting thickness of 0.05 mm and 0.1 mm
were employed for electropolishing in order to study the
effect of sample thickness and electropolishing time on
interface phase formation. Results suggest that the face
centered and the hexagonal interface phases in titanium
alloys arise out of processes occuring during the electroly-
tic preperation in thin foils [62]. The authors believe that
the fcc phase is formed in alloys containing low volume
fractions of beta and where the starting thickness of the
discs used for electrothinning is high (polishing times are
longer). The fcc phase was confirmed to be a hydride of
titanium. which forms as a result of an increase in hydrogen
concentration in the thin foils during electropolishing.
Westlake and Gray have shown the hydrogen content in
materials such as V. Ta. Nb can increase as much as four

times during electropolishing [63]. depending on the starting

41

thickness. The hexagonal interface phase occurs when the
amount of fcc phase formed is low and when the starting
thicknesses are low and/or the beta volume fraction is high.
It is suggested that the hexagonal interface forms due to
stresses at the alpha/beta interface arising from the thin

foil relaxation in the beta phase during electropolishing.

3. REVIEW OF RELEVANT ELECTRON-ANALYTICAL TECHNIQUES

Three relatively new techniques were used in phase
identification and solute partitioning in Titanium-6211.
1. Scanning transmission electron microscopy (STEM)

2. Microdiffraction in STEM
3. Microchemical analysis in STEM

Following is a brief review of the above mentioned techniques.

3.1 ScanninggTransmission Electron Microscopy

STEM served as a very important technique in the phase
identification and characterisation of the titanium alloy in
this project. The microscope utilised was a vacuum generator
model 5 installed scanning transmission electron microscope.
An understanding of the working. and capabilities of this
microscope is worth highlighting as part of this project.

In STEM an electron transparent Specimen is illuminated
by a small scanning electron beam. similar to that in the
scanning electron beam microscope (SEM) rather than by a
relatively large diameter station ary beam as in conventional
transmission electron microsc0py. STEM has all the
capabilities of TEM to define microstructure quantitatively
as well as to perform simultaneously chemical analysis of
areas of the Specimen on a tenth nanometer scale. Polymers
that get degraded under the electron beam in TEM can be
investigated readily in STEM for longer times. SEM
techniques such as imaging with secondary and primary back

#2

43
scattered electrons. production of channeling patterns etc.
can be done with STEM techniques; Fig (17) illustrates the
STEM set up. Incident beam is scanned over the specimen. and
the STEM image formation is detected for one image point at a
time. The detector is normally a phospher-photomultiplier
combination which provides an electrical image signal in
serial form for display on a cathode ray tube after recording
in an analog or digital form. To achieve good quality
imaging. a good signal to noise ratio is required which means
that the number of electrons scattered from each picture
element within a short period must be around 104 or more. To
attain high resolution imaging this number of electrons must
be concentrated within a very small probe size. This demands
a high intensity electron source from the use of a field

emission gun.—

3.1.1 Microdiffraction in STEM

An important feature of transmission electron microscopy
is the capability of obtaining both a high resolution
electron micrograph and an electron diffraction pattern from
the same selected area of the specimen. Consequently
considerable effort has been made to reduce the size of the
area from which the diffraction pattern is obtained. In
todays modern conventional transmission electron microscope
(CTEM). the standard selective area diffraction technique
(SAD) allows area selection of approximately 2000°A diameter

minimum. With the introduction of STEM, important new

techniquies have been developed which extend the capability

Electron
scurce

Condenser I

Condenser II

 

 

 
  
  
  
 
 
  
 

OOuOle deflection
scan cords

 

Secondary electron
oetecruon

Bactrscattereo
meetron detect-on

X-rav 000C100"

 

 

 

 

 

Protector tens to
comet detector
cottectron angle

 

lsansmutteo etectron
detectron (Ghent Fat-im—

.........

 

 

 

 

 

 

 

    

Screen
Annular ovum-tor Int ..1
scatterer! etectmns
than t ..«m ’
' Electron
energy toss
spOCItomctry

Fig.17.

45
of SAD to less than 30 A0. Using the criteria that the
minimum crystal size for a meaningful diffraction pattern
should be 5 times the unit cell size. the goal of minimum
area electron micro-diffraction has essentially been
achieved. The procedure used to obtain microdiffraction
patterns depend in part on the electron optical configuration
of the objective lens in the electron microscope. The most
straight forward technique uses the convergent beam
diffraction pattern (CBDP) obtained with the electron probe
formed for STEM imaging. Stopping the probe on the area of
interest forms a CBDP in the back focal plane of the objective
lens which may either be recorded on photographic film
directly or displayed on a CRT by means of Grigson coils
placed beneath the objective lens. The resolution of the
pattern is determined by the angular aperture of the
convergent beam probe which is typically about 10 mrad.
Defocussing the probe allows the angular resolution of
diffraction patterns to be improved to one mrad or better
from areas about 200 A0 in diameter. In another method the
Specimen is positioned approximatelymidway between the
objective lens pole pieces such that a near parallel incident
beam may be rocked about a point lying on the Specimen plane.
A sequence of dark field intensity maxima are formed at the
detector and subsequently displayed on the CRT. Diffraction
patterns from areas less than 30 A0 in diameter have been
obtained with an angular resolution of 3 mrad using this

technique. These techniques provide single crystal Spot

46
patterns from which phase identification and/or orientation
may usually be determined. An alternative approach employs
dynamical diffraction in CBDP to obtain more accurate data
from larger areas (100-500 A0) orienting the specimen on zone
axes where dynamical diffraction occurs.

A comparison of techniques shows that both the fixed
probe CBDP method and the beam rocking method yield
diffraction patterns from extremely small areas: however.
specimen contamination in existing commercial instruments
may be a serious problem in the case of the CBDP method.
suggesting that beam rocking may be more appropriate when
studying regions of less than 100 A0 diameter. Neither the
fixed probe CBDP method nor the rocking beam technique seem
to have deleterious effects on metallurgical Specimens with
regard to beam heating or radiation damage at accelerating
voltages up to 100 KEV. However an intense beam of electrons
is required to obtain information from such small areas.
hence there may be damage in many of the non metallic beam
sensitive materials. Other methods of electron diffraction
are the use of optical means to obtain selected area
diffraction patterns and high diSpersion or small angle

electron diffraction.

3.1.2 Electron Beam Microanalysis in STEM

Electron beam microanalysis is a powerful analytical
technique capable of performing elemental analysis of
microvolumes. typically on the order of a few cubic microns

in thick samples and considerably less in thin sections.

1+7

Analysis of X-rays emitted from a sample can be accomplished
by energy dispersive spectrometers which permit analysis by
discriminating among X-ray energies or by crystal spectro-
meters which use a diffracting crystal to select the
wavelength of interest. The former method is called energy
dispersive Spectroscopy (EDS) and the latter wavelength
dispersive spectroscopy (WDS). is an older method. EDS and
WDS can analyse elements begining with sodium in the periodic
table. It is possible to detect less than 10'16 grams of

an element present in a Specific microvolume of a sample.

The development of the EDS system has been strongly linked to
advances in computer technology. Electron energy loss. and
wavelength dispersive spectrometers are routinely interfaced
to the EDS computer for data collection. storage and

processing.

3.1.3 Principles of EDS system operation

The backbone of the EDS system is a silicon diode
typically 12.5 mm2 in active area and 3 mm in thickness. If
the diode is reverse biased by approximately 1000 volts a
depletion region is established. All of the normally present
electrons and holes in the depletion region are removed (by
the external field). If an X-ray photon is absorbed within
the depletion region. producing a number of electron-hole
pairs these charge carriers will be swept out by the electri-
cal field and they will appear as a pulse of charge on
opposite sides of the diode. This charge is converted into

a voltage pulse. amplified and shaped. its amplitude is

#8
converted into a digital signal and the resulting information
is analysed.

The total time the system is not available to accept
pulses (during the processing of a pulse or while the system
is gated off due to pile up relection) is called "dead time .
Since pulses may have been arriving while the ADC was busy,
the time of the analysis is extended by the live-time correc-
tor to compensate for this dead time. Live time correction
is necessary to measure the true count rate for a given
element. a parameter required for quantitative analysis and
for accurate comparison of two spectra accumulated at differ—
ent count rates. Because the electronics of the EDS system
must have both sort and count X-ray events. the total count
rates that can be processed are much lower than those of a
WDS system (where the sorting is done by crystal diffraction).
The resolution of a given ED Spectrometer is defined as the
full width at half maximum height (FWHM) of the peak above
background at 5.9 Kev (Mn KG). The resolution is determined
by statistics of charge production. detector and preamplifier

noise. and degree of incomplete charge collection.

3.1.4 Qualitative EDS microanalysis

Perfect qualitative analysis would permit identifica-
tion of all elements present in any selected region of the
sample imaged on the microscope. Limitations on the ability
to perform such an analysis arise from considerations of

detectability limits. Spatial resolution. and spectral

resolution.

The minimum detectable quantity of a given element. or

49

PERCENT EFFICIENCY

 

 

‘8 d 1 114111d dqqqd

      

 

 

I 11
re 1
I 1
I 1
T ..
r I.
.. 920.053. .0. m: ...
5336203"...
5 t
I l
r l
I l
1 1
. l
t l
r l
r l
I l
.o .p Pl. .PLLg . _ p _. .p.— p p kph b...
O; ..o .0

$35 m2mmo< .z .6<
02606. 625532 .9 w 33 9:8? mm 2:592 3.0.253 Nw t3

mwm.~m.

 

50
its detectability limit varies with many factors like the
X-ray line used for analysis. accelerating voltage. matrix.
detector efficiency. sample geometry. peak counting time. and
peak and background counting rates. Elements below fluorine
in atomic number are generally not detectable at all.

The silicon diode spectrometer is nearly 100 % efficient
over a wide portion of the analytical range. At low energies.
however. the efficiency goes down because of absorption of
X-rays in the beryllium window and other detector surfaces.

At higher energies the spectrometer again becomes less effi-
cient due to the ability of these X-rays to pass completely
through the detector. Fig.(18) illustrates this.

Spatial resolution plays a very important role in micro—
analysis and data collection by the EDS. The goal in electron
beam microanalysis is to analyse small volumes of a sample
that correspond to a particular structural feature in the
electron image. Resolution of 100 A0 is possible with thin
sample sections and small intense probes. The main consider-
ation in the weakening X—ray Spatial resolution is the spread
of the electron beam as it penetraufisthe material. Secondary
electrons which form the image are very low energy, coming
from onlyttthin surface layer on the order of 100 A0 in depth,
where the beam has not spread significantly in the lateral
direction. Secondaries continue to be generated as the elect-
ron beam penetrates the sample. but do not escape and do not
contribute to the image. High energy back scattered electrons
can escape from much greater depths in the sample. thus a back

scattered image is of poorer resolution than the secondary

51

 

 

 

Fig 19
. . Esca
pe de
plh for
secondary electr
on,back
scafler
ed

eflecl
ron '
emission and X
Jay

52

image. Finally. X-ray can contribute to be generated by the
primary beam and being able to travel further in matter than
electrons. escape from still larger depths in the sample.
Fig.(19) illustrates this fact. It is clear that one can see
features in the secondary electron image that can not be
uniquely analysed in the X-ray analysis. For example a
surface particle 1 um in diameter but only 100 A0 thick. may
be readily visible in the image. however. most of the X—ray
information comes from a region below the particle and the
analysis obtained will not be specific to the imaged particle.
Hence one must be aware of this fact to interpret microanalysis
results. One technique to improve X-ray spatial resolution is
to decrease the accelerating voltage thereby decreasing
penetration of the electron beam in the sample. By using thin
sections the electron beam is not given the opportunity to
Spread very far in the lateral direction. In this case the
resolution continues to be improved by reducing the probe
size.

Spectral resolution is important in qualitative analysis
because the X-ray lines of one element may be hidden under
the lines of another causing the analyst to miss the identi-
fication of the overlapped peaks. In EDS peak overlap
situations like the Mo L and S K lines can only be rectified
by use of a crystal spectrometer.

Interpretation of SEM. TEM, X-ray diffraction techniques

for this project will be discussed in the results and discus-

sion section.

4. EXPERIMENTAL PROCEDURE

4.1 Heat Treatment

 

Ti-6211 plate (as received) was cut into smaller blocks
of 2 cm by 1 cm in size using an abrasive cut off wheel.
Samples were then polished to remove sharp projections and
then encapsulated in a quartz tube at a vacuum of 10'5 torr.
Solution treatment was done at 1050°C and then all the
samples were step quenched at progressively lower tempera-
tures of isothermal holding followed by quenching to room
temperature. One particular sample was step quenched to
450°C and held for 72 hours to study the ordered a2 phase
formation. Heat treatment schedule adopted is illustrated

in fig (20).
4.2 Study of Heat Treated Samples

4.2.1 X-ray Diffraction

Heat treated samples were cut to fit into the Specimen
holder of the X-ray diffraction unit. Sample surfaces were
polished and lightly etched before subjecting it to
diffraction. A philips Norelco X-ray diffractometer fitted
with a LiF crystal Monochrometer was used with Cu Ka

radiation.

4.2.2 Scanning Electron Microscopy
Heat treated samples were polished on 240. 320. 400 and
600 emery papers.'followed by disc polishing using 0.05

micron alumina. Samples were then etched for 60 seconds with

53

54

 

[HEAT TREATMENT—l

 

I

l. ' LP sownorttzso (to'So'c)]

 

 

 

 

60“: 6 0": 7|O'C 750‘C BOO'C 900’C 9 O‘C

 

 

IjUENCl-IED T0 R.T. IN HATE?!

 

 

 

*—_'E_ SLOW COOLED T0 KT]

Fig . 20. Heat treatment schedule

55
Krolls reagent comprising of 95 cc H20. 3.5 cc HNOB. and
1.5 cc HF and then washed with water and methanol. A Hitachi

5-415A scanning Electron Microsc0pe was used for this study.

4.2.3 Sample Preparation for TEM/STEM
The preparation of samples for TEM and STEM studies were
identical. comprising of the following steps.
1) Specimen slicing to about 0.5 mm by diamond wafering
blade.
2) Hand grinding on coarse emery paper to a starting thick-
ness of about 1501Jm.
3) Shaping the samples to 3 mm diameter and 1507nm thickness
4) Jet polishing to achieve a dish on either side of the
sample using the following chemicals and conditions.
30 ml perchloric acid
175 mls butoxyethanol
295 mls ethanol
voltage 120; current 130-140 mA
5) Final polishing to achieve a small hole at the center of
the specimen.
6 ml perchloric acid
175 ml butoxyethanol
295 ml ethanol
voltage 12-50 volts. current 20mA. Temp -45°C
to -55°C

5. RESULTS AND DISCUSSION

5.1 X-ray diffraction studies

All the heat treated samples. representing various
isothermal holding temperatures. were subjected to X-ray
diffraction. Figs (21) and (22) illustrate the critical
bragg angles at which the phases are identified. Between the
20 values of 330 to 43° and 720 to 800 identifiable primary
or untransformed beta phase and the orthorhombic martensite
phase have been detected. The (202)a" martensitic peak shows
a decrease in intensity for holding temperatures from 650°C
to 900°C. and decreases in intensity for holding temperatures
greater than 900°C. This suggests the fact that there is a
critical temperature range in which orthorhombic martensite
is formed. This result however is affected by a strong
texture and grain growth. There is also a shift of (110)B
peak to higher angles with decreasing heat treatment
temperatures. This suggests that there is a contraction in
lattice parameter of the retained beta phase with decreasing
temperature. Molybdenum is the only element in Ti-6211 which
decreases the lattice parameter of the beta phase and hence
it is suggested that supersaturation of the beta phase with
Mo occurs at lower annealing temperatures. This could lead
to a critical beta phase lattice parameter at which B-a"
transfomation occurs. From the model proposed by Mukherjee

Kato [27] and from X-ray diffraction experiments lattice
56

57’

(2051)

Ouenched
80' | 72'

  

I too'C
sso’c

 

I
1‘ Furnace Cooled
l 72'

 

1718.21. Composite X-ray peaks; CuKQ radiation

58

'5.
e,
’8‘

9.

2‘
O
15:

 

 

 

b

(CC

E

3

8. a}

sec

...—{Furnace 090'0‘

§

33'

fLLLLEELEL
ii if
. at?

Fig.22. Composite X—ray peaks; CuKa radiation,

59
parameter of the orthorhombic martensite phase was calculated
as a = 3.04 A0. b = 4.94 A°. c = 4.64 A0.
The lattice parameter of the beta phase at room temperature
was also calculated to be a = 3.23 A0. Since the MS
temperature for the a"-martensite is higher than room
temperature. the actual lattice parameter of the beta phase
at the MS temperature is greater than 3.23 A0 due to thermal
expansion as well as due to lower Mo content of the beta
phase at higher temperature. However. in the first
approximation. these lattice parameter values can be used to
compute the habit plane of the a" martensite by using the
model suggested by Mukherjee and Kato 27 . The calculated

habit plane from such a calculation is: (239).

5.2 Transmission Electron Microscppy

TEM studies were conducted to investigate the phases
present at various isothermal holding temperatures. Before
discussing the results obtained it would be appropriate to
refer to the schematic beta isomorphus diagram of Ti-6211 as
suggested by Williams Fig (23). .As the isothermal holding
temperature decreases the volume percent of beta decreases
(lever rule). and the alloying-element content in the beta
phase increases rapidly. The alloying element in the alpha
phase increases slightly with decrease in isothermal holding
temperature. Thus depending on the solution heating
temperature. the beta phase in Ti-6211 alloy can undergo
transformation to a' or a" martensites or can be retained in

a metastable condition upon quenching from elevated

60

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a" RETAINED B ——-—
954°C
(l750°F)
87l°C
(l600‘F)
B
788°C.
(I450°F)
Transformed
'0 all
Transformed \.
to (2’ \\\ V
C! ' Remoined
\ \V as B
\ ‘ -
a + B \ \
\ \
\ .
1 \
“'62" °/. SOLUTE —>

Fig.23. The schematic phase diagram of Ti 6211

61

temperatures. At high n+8 solution treatment temperatures.
beta transforms to o'. at the intermediate temperature range
beta transformed to a" and at low temperature range the beta
phase is retained. The transformation products depend on
solute concentration of the beta phase.

TEM studies for 950°C isothermal holding temperature
reveal twinned hexagonal martensite formation Fig (24) and
the associated selective area diffraction patterns confirm
the hexagonal crystal structure. Bright fields for 850°C
isothermal holding temperature show alpha platelets with the
interface phase exhibiting high dislocation density fig (25)
unlike the 950°C specimen which does not reveal any
predominant interface phase region. The alpha platelets are
of the Widmanstatten type with dislocation lines running
across them. Examination of 750°C TEM Specimens reveal a
more pronounced Widmanstatten morphology of alpha platelets
with dislocation fields surrounding the interface fig (26).
suggesting the fact that the interface region is stressed as
a result of alpha transformation from parent beta phase.
Specimens held at 700°C isothermal holding temperature Show
beta platelets fig (27) as confirmed by SAD patterns. TEM
results Show that as one goes to lower isothermal holding
temperature the tendency for beta to remain untransformed is
more. and the interface phase ceases to form for temperatures
below 750°C. The temperature range of 750°C-850°C allows the
interface phase to be more stable. According to Willams the

interface phase in the alloy is stable between 8000 and

62

 

950° C SAMPLE

Fig.24 Bright Field and associated selected area diffraction

63

 

Fig.25. TEM sample isothermally held at 850°C showing dense

dislocations at the interface

 

TEM sample isothermally held at 750°C for one hour

showing alpha plates.

 

64

 

e
ZONE AXIS [T23]b.c.c
, a/b=1.l
/ c/a=l.38
(2;) c/b=l.515
6
(5;. .\ 4° "' ’5
3) (0‘6‘
0 ”/6 e
4; :EthE) 0
:0 ,‘/ eke
e ./ 019 0
d 0

750°C SAMPLE

Fig.26 Bright Field and associated selected area diffraction

showing alpha plates at the interface

 

 

65

 

 

0 AT
A L‘ ,1'1 ‘7‘
e , ,—
.' I ‘
° 6% _ 4 if
.4, 'n *23
\’K ‘2 ‘- .
6‘2) p :7 6 /
. ;,<5;
i in
0,“

700° C SAMPLE

Fig.27

ZONE AXIS [123] b.c.c
b/a=1.032
b/c=1.379
a/c=1.310

66

 

Fig.28. TEM micrograph of sample held at 850°C for one hour

showing dense striations at the interface.

67

850°C. Dense internal structure fig (28) at the interface
region suggests an hcp structure as per Williams. However
since the appanance of the interface phase is very sensitive
to contrast conditions, no authoritative comment can be made

in this regard.

5.2 Scanning Electron Microscopy

Various isothermally held Specimens were observed under
scanning electron micrOscope. All the SEM micrographs
reveal beta network interspaced with alpha. The amount of
beta phase or titanium martensite decreases with decreasing
temperature as seen in the SEM micrographs. In the case of
the annealed sample. the n+8 network is more widely Spaced
owing to the long duration of anneal permitting diffusion of
atoms to longer ranges. The ordered structure 02 could not
be resolved in SEM owing to its limitation in magnification.

Isothermal annealing at 850°C reflects a higher
percentage of beta transformed product as compared to
temperatures below it. Figures 29-35 illustrate details
Fig 35 reveals voids present in the alloy. The acicular
morphology of martensite is clearly revealed with twins

running across.

5.3 Scanning Transmission Electron Microscopy

The primary intention of using the technique of STEM was
to study solute partitioning tendencies in this alloy at
various temperatures. Microdiffraction was also done
especially for determining the ordered phase a2. Appropriate

bright field pictures were taken to illustrate the morphology

 

Fig.29. SEM micrograph of sample isothermally held at 550°C
for one hour showing beta stringers interspaced with

alpha

/

e. K’ . . I
Betas-25K "6"2t1'fi

Fig.30. SEM micrograph of sample isothermally held—at 650°C

for one hour showing beta network interspaced with

alpha.

- -\. edez, 25K'715*Wfi

Fig.3l. SEM micrograph of sample isothermally held at 850°C

 

for one hour showing alpha at grain boundaries.

 

Fig.32

70

“ ';’
e, '

’3331 25K 38 2N" ‘

 

Fig.33. SEM micrograph of sample isothermally held at 700°C
for one hour showing beta network interspaced with

alpha.

ANNEALED

 

Fig.34. SEM micrograph of annealed sample held isothermally
at 450°C for 72 hours showing beta stringers

interspaced with alpha.

- seas 25KBB=I=2N

 

Fig.35. SEM micrograph of sample held isothermally at 900°C

for one hour showing voids and acicular martensite.

   

STEM nticrographrof sample ‘17
held at 750°C for one hour
showing butterfly martensite.

Fig.36.

72

 

Fig.37. SI‘EM micrograph of sample

isothemally held at 750°C
for one hour showing

acicular martensite.

Fig.38. STEM micrograph of sample held
isothemally at 750°C for one

Wins in martensite.

 

(00-02600

3°00 4»

[0000

 

 

73'

--t

71621! 750C

LT-IIO SECS

2.000 4.000 0.000 3.000 10.000
curacy (KEV)

Fig.40. Hicrochemistry of 750°C specimen

“

74

 

held at 750°C for one hour
twins in nartatsite.

Fig.39. SI‘EM micrograph of sample isothemally
shwihg

 

Fig.41. STEM micrograph of sample
isothermally held at 650°C
for one hour smwing beta.

75
of phases present at certain temperatures.

Fig 36 reveals a typical butterfly martensite Showing
two variants formed after 750°C isothermal treatment. This
morphology of martensite has also been obServed by other
workers in this field. Fig 37 reveals acicular martensite
with signs of twinning (temp 750°C). Figure 38. 39 clearly
Show the twinned region within the martensite. Results
indicate that isothermal treatment at 750°C exhibit two
morphologies of martensite as discussed. Solution partition-
ing tendencies at this temperature was studied by utilising
energy dispersive analysis at a region within the martensitic
plate as indicated. The EDS is taken for the plate shown in
figifi3. Fig 40 corresponding to fig 39 shows that all the
beta stabilising elements Ta. Mo. Nb are concentrated in this
region with a comparitively low count of aluminum which is an
alpha stabiliser but nevertheless is present in this alloy
with a maximum weight percent constituent (6%th). Fig 41
reveals an untransformed beta region. solution treated at
650°C and fig 42 reveals the microchemistry from a region
within the plate. It is seen here that the partitioning of
Nb. Mo. and Ta predominantly to beta phase increased as
isothermal annealing temperature drops down to 650°C from
750°C. In general the diffusion rates of aluminum in beta
is more sluggish than in alpha. because aluminum preferentially
dislolves in the latter. Similarly the beta stabilising
elements Mo. Nb. and Ta dissolve preferentially in the beta

phase and thehfdiffusion rates are higher in this phase.

Ul-SZCOO

76

3.1.1» ~ — '-
1509-0
l 0 0
4»
1°..-0 . .0
> . I - T
A
4D A x . . .
1* L ”H e' e r
. l .0 i o e ‘ t
t '1 . uqyflfivf g . .0 7
. ' N“ . . . O A
0 0'. S 0 r
5.0-0 I. ...-0’ ‘I o .. I . 0 .
.. * N X’.‘0:. I, q ,0 I:
l ... x .A; J “.0 ‘ 00
o . J...
f
{D .
0 3
.

 

 

‘ AA A ILL

' trrfirrtIrrrtttttrltrtlrtrrrittlrtrrrttrrrttrwrrtr

..... .... 4.... 6.... ....I "....
ENERGY (KEV)

Fig.42. Energy Dispersive Analysis
of 650°C sample

77

 

Fig.43. STEM micrograph of sample isothermally
held at 8(I)°C for one hour showing
alpta plates and beta interface.

 

Fig.44. SIEM micrograph of sample
isothermally held at 8(D°C
for one hour shaving alpha
plates.

78

As predicted in the schematic phase diagram of 6211 the
solute concentration of beta phase should have a relatively
wider range as compared to alpha phase. as one decreases the
isothermal holding temperature. Figs 43. 44 reveal alpha
platelets with an untransformed interface phase. Energy
dispersive analysis was done at a region within the plate and
also at a region near the interface as shown by the markers in
Figs 43. 44. Fig 47. and fig 48 reflect microchemistries
from a region within the plate and at the interface f
respectively. Within the plate it is seen that beta
stabilising elements Mo. Nb. Ta are present at a much lesser
concentration than aluminum whose counts are prominent. As
we near the interface we find a drop in the concentration of
aluminum. with the concentration of beta stabilising elements
more or less being the same. This means that the interface
is a region which rejects aluminum and encourages the
dissolution of beta stabilisingelements. There is a strong
possibility of the interface being untransformed beta.

The.same analysis was done for a specimen which was held
isothermally at 700°C. Figs 49. 50 illustrate the regions
where the analysis was done. Figures 51. 52 Show the
chemistries of the regions present. It is seen that there
is a strong tantallum concentration as compared to the other
alloying elements at this temperature within the beta
platelets. The morphology of the platelets and solute
concentration indicate the presence of beta region. and as we

go towards the interface there is a drop in the tantallum and

0542600

79

 

Lr-ton sacs i i rtezlt a0uc
2000-4 —""
J. .
A .
o L
1500-« ' -
b
o .
1...-" o .
1' 3 °
0 o .A l .
0 . ' . . 0 e .
N O 0
$0 a. O: . ‘0
5.9-” o .0 d‘ g 3 “‘0“
e 0...... ,.. 'J':'.e' .
4r 0 O -
' O .
o ... ‘1
f 3:: ‘: x T
4. - e {0. : ‘0‘.
' 53 ‘ ‘0'. .0. \\00 e : A
v ' ’ ‘ae‘?s~&r-- “ ..
0 ’0 00 “no. W..::.w.
‘2
“- r-t-t-r-r-t-t-r-t-r-r-r-t-t-t-t-H-r-r1-1-t-t-t-r-t-t-t-t-1-r
3.099 2.390 4.29. 5.00. 3.6.9 10.nec

ENERGY (KEV)

Fig.47. Energy Dispersive Analysis
of 800°C sample.

 

80

 

 

 

L't'=ll3l‘ 'JL'Cf, l l .\t( (LI-fl L‘NlA L/JLL
—_
tun-L
4)
o
(h
500-» .
T
l.
0 -
o
I I
000-» ,
A I
L
d
O
l .,
O
0 o
:0
0 °
C
500-0 . - '
a .fi o
0} . .0 . '.. .
sn- ° '
0 a.‘ BS .{ . ed
0 M". e. a?“ .
fin
‘I '5 .O‘M ‘
0 0‘ .... T
" 3 t g: A T
O . . . 0.
1L .. ' W Mn”.
0 0 ”~09“th
- A AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
' IIUUITUrIFITFIFUIIrlllITIIIITIT1IIIIIIIIIUWIIVII1
....I 2.... 4..°. 6.... 3.... l.....

IIIRGY (KEV)

1 ..IIIKIV LTI 1..8£CS. LABSL:EXBC(20H) DATA LABEL
1.24.: 1.7GIKIV GROSS- 123.. III?- 4580

4.2..: 5.25.KEV GROSS-I 32727. NET- 3123.2

Fig.48. Energy Dispersive Analysis
of 800°C sample from the
interface region.

Iii-.2600

81

 

 

 

 

- .. . T ‘
Lr'lzfl uCCb I f '3'15211 800C
zone-l —-
D ..
L .
A .
a L
. o
1500-0 ' -
0
e O
0
o f
0
4L .
.
1‘..-" e t
0
1b .
.. .
0
.-
‘T ’. '
h A . O
'0‘ t
a O ' . ‘ .
I .\ u o ‘0.
e o a" $ . '
5"” °‘ ' ‘30.: ‘- 3 ..u-x'
.0 O O. 0.
* ; 1 ~ '
1 °’ 1&3 ‘f’e 3f . .
i I. e 00 ° ‘1
.,, Mauv' . *5v
o .’i “0 $0.
I
..~ 3} 6':
t if
1 .
a _~ _ ._. l l A
- IUIrIIIIIIIIIITl]Ullrl‘r'Trr‘UIUU—IltiirlUlrITIWII
0.0.. 2.0“0 4.05” .905 3.00. 113.97.?

cursor (nevi

Fig.47 and Fig.68 superimposed.

82

 

 

Fig.49.700°Csample
showing region of
analysis

 

Fig.50. 700°C sample
showing interface
of analysis

83

—4

LT-IUH SECS T1621! 700C

2090-?

)4

....-i

0
0°. *9.

 

 

 

c . "
o t) . .
u o
n 1000-0 .
1' ° .
s 4}
1> .
4)
500-0 I , . .
1 . .
t. ‘ . -‘ I
: 8 :‘ ' :
o u . ‘ U
40 . o B
. "o; '- .
‘; <3! 5 0 g ‘
0 . O 0 v
‘3' ’pr‘fa'f’fz'n ‘ I I
u . ““~"' on g .
r, ' ' 0n WWW-anavM-e
9 r 134% rfiri YES! I ll Trri T’TT I 11 r’rr 1 I1 I It r‘t1 l r1 T‘t1 t It I It I It ll
0.000 2.000 4.000 6.000 0.00c tu.coa
ENERGY (<EV)
hfil‘ 0 (12). l ..OZOKEV LT- H‘OSECS. LAutLLz't'162ll 766C
REG 0 tool 1.240: 1.700xev uuoss- 15155 NET- -u§05
Rec. l lull 0.209: 5.269uuv CROSS: tnvflnl NhT- 91452
ltl-ZC 2 [02) 2.!70: 2.44mftv :;.tttf;t: 1'15) anl‘s l1)?
REG 1 lull 1.32": 1.94CLLV anus": [369% nt:- Irut?

Fig.31. Energy Dispersive Analysis of 700°C specimen

84

T
11-100 secs 1 :16211 700C luTuRrAc
2000-0 '-
r»
1500-0 .
T
4)
0 T 0
c A
o {D .
u a
n 1000—0 I
r ' .
s P

 

l' '
r ’ ’
s..-" . . . e
t. _: H - ..
I ' NOra . ' '
0 ‘ t C " L e e
' vet; . . .
‘ ‘$; ' ”ppfl.w 0.". I z
’ ’H9-‘fl‘~‘V"‘9“fii’"’n—Aau-nq¢a¢—uhpn
"‘ -T-r1-r-r1-T-r1-1-r1-1-r1-r-r1-r-r1-t-r1-1-r1-r-rfl-r-r1-r-r1-r-r1-t1
3.00. 2.990 4.002 6..fll 8.00. 10.000

ENERGY (KEV)

£22). I 0.122.11t2V LT- IOOSECS. ”95514716211 708C INI'ERFAC
0 ID.) 1.240: 1.140.131 CROSSI- 9197 dé'l'l -S.é‘:
I 101) 0.20.: S.263.’.LV Col-355- 59949 NETI 5(0):
2 1021 2.123: 2.4445LV SHUSJ? 2546 art: 214
1 11211 1.5711: 1.901‘-r;v -’;.lto:.:.= 1:111!) 42".- 7'15?

Fig.52. Energy Dispersive Analysxs of specimen at
interface. 700°C specimen.

85
aluminum concentration. The rejection of aluminum at the
interface is consistent with the solute partitioning study
done with the 800°C specimen. Studies indicate that in the
tmeperature range 6500-70000 there is a greater enrichment of
tantallum in the beta region because of increased diffusion

rates of tantallum in this temperature range.

5.3.1 Analyses of Ordering in Ti-6211

As discussed before. the phenomenon of ordering was
investigated in this alloy using STEM techniques.
Conventional TEM samples were used for analysis. Prior to
making TEM samples. the alloy was subjected to 72 hours of
aging at 450°C. Fig 53 reveals a typical alpha platelet with
02 ordered precipitates. Fig 54 shows a magnified image of
the 02 ordered precipitates. It is clear that the precipi-
tates are extremely fine i.e. less than 0.02 microns and can
only be resolved under very high magnifications like
500.000 X. The precipitates are not strictly elipsoidal but
vary in Size and morphology depending on local changes in
misfit strain energy withinthe alpha mattrix. Solute
partitioning tendencies were studied in the ordered region and
in a region other than the ordered precipitates signifying
the average compositioncfi'the alloy. Figure 55 and fig. 56
reveal the microchemistry from a precipitate zone and an
average zone respectively. The precipitate zone registered
an aluminum count about 2.2 times that of the average

mattrix count. This is consistent with the fact that in the

ordered structure (DO19 lattice) of composition Ti Al,

3

    

Lé‘

Fig.53. SIEM micrograph of sample annealed at AESC for 72 hours
showing ordered precipitates.

 

Fig.54 SI‘EM micrograph showing magnified image of ordered alpha2 precipitates.

tat-03:00

87

 

T
Lr-100 secs 1 exacrz-w) 0070 LABEL
boo 0
2000
‘00 J)
C
N .
'r ‘38 L ,
..A. 0. o ’- . . T
e ...“. of. .. . . o .. A
. .. ..o~'..o~.. . .J'. - .
°-' ‘t'r-r-r-r-t-r-t-t-«t—r-r1-t-r-4-t-r-r-t-t-1-é-1'444 4:14:51" 1-4-‘1-{04'44-44"?3’1
..OI. 2.19111) 4.0011 6. 0111' I. 605‘ 11!. 0110

LJERCY (KEV)

Fig. 55. Energy Dispersive Analysis from
ordered precipitate region(a1pha2 )

m-010:0t1

88

 

T
LT'I.‘ SECS I N“! (3)
-
{b
5.6-0
Q
C
(b
O
0
O
O
. O
. O
A T
LJ\ N33 - _ r
.§.0 .8 «0.0 . ~ . A A
' ’hvaflbufhheedfif%:?fi~‘?fif ' 0 g,
-1 -t-1-t-t-t-t 1444-14-
0.60“ 2.flfll 4.0“” 6.00" 3.00“ [0.990

nuxuur (nevi

Fig.56. Energy Dispersive Analysis from region around alpha2

89

 

. Ordered
«\ I «\’L\ ._(
,\ \\O’ 3030
\‘\ \
~ ~. \
(Q‘ \ 01‘1’ . 201.0
C» \ 7'7-
?1, \ 22
// FWO
\ ‘ w 6

ZONE AXIS [2423]h.c.p
a=2.35

b 3.0
cl4

ANNEALED SAMPLE

Fig.57. STEM picture showing microdiffraction fran ordered alpha2 precipitate zone.

90
aluminum occupies 25% of the total atomic concentration. as
compared to the disordered alloy with an average random
aluminum concentration of 10.4a at %. Microdiffraction in
the precipitate region fig 57 revealed fundamental hcp Spots
along with superlattice Spots i.e. spots which were not
allowed in a hop lattice according to the extinction rule.
This further confirmed the a2 ordered structure in this alloy.
Further investigation of the a2 structure is in progress

using differential scanning calorimetric techniques (DSC

5.4 Summary and Conclusions

 

1. TEM studies show that hexagonal martensite forms on
quenching from higher isothermal holding temperatures of
900-9500C. and on progressively going down in the
isothermal holding temperatures, there is a tendency for
beta to remain untransformed.

2. As isothermal holding temperature decreases the morphology
of alpha platelets assume a Widmanstatten shape.

3. The amount of equlibrium alpha increases at lower
isothermal holding temperatures.

4. The interface phase is stable in the 750°C-850°C range.

5. Solute partitioning tendencies in this alloy indicate a
greater concentration of beta stabilising elements at
lower isothermal holding temperatures. The temperature
range 650°C-700°C encourages a greater tantallum
enrichment in the beta phase, because of increased

diffusion rates of tantallum.

‘ T 1m _.-'.

:‘
<

91
There is a marked drOp in the concentration of beta
stabilising elements as one goes away from the interface
region towards the center of the transformed beta
product, indicating that the interface region has the
composition of untransformed beta.
Ti-6211 exhibits ordering at QSOOC aging, leading to fine

ellipsoidal a2 precipitats of Ti Al composition and D01

3

structure. Further investigation is in progress using

9

Differential Scanning Calorimeter (D.S.C.) techniques.

10.

11.
12.

13.
1a.

15.

16.
17.

REFERENCES

Hammond. C. and Kelly. P. M.. Acta Metall. 11.
P 859. (1969)-

Williams. J. c. 2nd Ti-lOO review, NRL-March, (1981).
Murakami. Yotaro. Titamum'80 Science and Technology

Frost F. D.. Parris, W. M. Hirsch. L. L.. Doig. J. R. and
Schnartz TRANS. AIME 5g. p 231. (195“).

Brotlen. F. r., Harmon. E. L. and Troina. A. R., Journal
of Metals. Z, No 4, Sect 2, p #13. (1955).

Bohm, H. and Westphal "Zeitschrift fur Metallkunde" B. 47
No.8, 8. 558. (1956).

Bohm. H.. Westphal "Zeitschrift fur Metalkunde" B. #7.
NO. 8. So 558' (1956).

Williams. A. J.. Cahn, R. w. and Barrett, Acta Met, g,
117 (1959).

Williams J. C., Polonis, D. H. and Taggart, R., Ref 3,
P 733-

Mcmillan. J. C.. Taggartir and Polovis. D. H., Trans.
(Met. Soc) AIME. 232. 739 (1967).

Gamet. P. and Christian J. w.. Acta. Met 1, 534, (1959).

Armitage, w. K., Physical Props of Martensite and
Bainite Iron and Steel Institute Report #93, Condon 1965.

Williams. J. C. and Rhodes, C G.

Williams J. C. and Hickman. B. S. Met. Trans 1, 16h8
(1970)- ‘

Nishiyama. z.. Sato 8., Oka. M.. Trans Japan Inst,1v1et.. g.
127 (1967).

Banerjee. S. and Krishnan. R. Acta Met, 12. 1317. (1971).
Williams. J. C. and Rhodes, C. G.

92

93

 

 

REFERENCES

18. Blackburn, M. J. and Feeney, J. A.. J. INST Met, 99, 132
(1971).

19. Blackburn. M. J. and Williams. J. C.. Trans. (Met. Soc)
AIME, 242, 2461 (1968).

20. Williams. J. C.. Rhodes. C. C.. Manuscript in Preperation

21. Williams, J. C.. "Titanium Science and Technology"

22. William. J. C. and Blackburn. M. I., Quart. Trans. ASMi
6OI 323.

23. Crossley. F. A. and Lindberg, Proc. of ICSMA...2 3.
ASM. 841. (1970).

24. Oka. M. and Taniguich. Y., "Titanium 80 Science and
Technology" AIME, 1. p 707 (1980).

25. Sasano. H. et al., ibid.. p 717.

26. Burgers. w. G. Physica IS GRAV. 1, p 561. (1931).

27. Mukherjee. Kato.

28. Sircar. Krishnan, Mukherjee, "Phase transformation in
6211" ASM-AIME Fall Metting 1983.

29. Silcock. J. M., Davies, M. H., and Hardy, H. I., Nature.
175. 731 (1955)-

30. Frost. P. D.. Parris. w. M., Hirsch. Jr. L. L., Doig, T.
R. and Schwartz. C. M., Trans. ASM 46, 231 (1954).

31. Silcock. J. M., Acta Met.. 6, 481 (1958).

32. Bagariatskii. I. A. and Nosova. G. I., Sov. PHYS. CRYSTALL..
lo 15 (1959)-

33. de Fontaine, D.. Paton. N.. and Williams. J. C.. Acta
Met.. 12. 1153 (1971).

34. de Fontaine. D.. Acta Met.. 18. 275 (1970).

35. Sass, L. L., Acta Met.. 12. 813 (1969).

36. Frost, P. D.. Parris. W. M., Hirsch. Tr. L. L., Doig,

 

94

REFERENCES

37.
38.

39.

40.

41.

42.
43.

44.

as.
#6.
47.

48.

49.
so.
51.

52.

53-
54.

J. R. and Schwartz, C. M., TRANS. ASM, fig, 231 (1954).
Hickman, B. S., TRANS. AIME, gag. 1329 (1969).

Blackburn, M. J. and Williams, J. C.. TRANS (Met Soc)
AIME, 242, 2461 (1968).

Brammer, W. G. and Rhodes, C.G., Phil. Mag., 16. 477.
(1967)- '

Brammer, W. G. and Rhodes, C. G., Phil. Mag., 16, 477,
(1967)-

Blackburn, M. J. and Williams, J. C. Trans. (Metsoc)
AIME, 242, 2461 (1968).

Hickman, B. S., TRANS. AIME, 245, 1329 (1969).

Williams, J. C.. Hickman, B. S., and Leslie, D. H., Met.
Trans., g. 1913, (1971).

Williams, J. C.. Paton. M. E., and Rhodes, C. G.,
manuscript in preperation.

Williams, J. C. and Rhodes, C. G.
Williams, J. C. and Rhodes, C. G.

Aaranson H. I., Decomposition of Austenite by Diffusional
Processes, Zackay and Aaranson, eds., Interscience 1962,
p 387.

Williams J. C., Paton N. E., and Rhodes, C. G.,
manuscript in preperation.

Kauffman, L. and Bernstein, H., Ref. 3, p 361.
Blackburn, M. J., TRANS. AIME 222, 1200, (1967).

Williams J. C. and Blackburn, M. J., Ordered Alloys
Claitors Press, p 425, (1970).

Crossley, F. A. TRANS. AIME, 242, 726, (1968).
Blackburn, M. J., TRANS. AIME, 242, 728, (1968).
Eillismd, J. C. and Blackburn, M. J., unpublished

95

REFERENCES

55-

56.

57.

58.

59.

60.

61.

62.

63.

reasearch, (1973).

T. K. G. Namboodri, "Ti-Al phase diagram", Materials
Science and Engineering, 51, p 21-22, (1983).

Proceedings of Ti-6211 Basic Research Program August
(1982), edited by B. A. Macdonald, 0. P. Arora, B. B.
Rath, R. D. Shull, et al. on Phase stability.

C. G. Rhodes and J. C. Williams: Met Trans A, 1975,
6 A, p 1670.

Harold Margolin, Ernest Levine, and Morrisyoung, Met.
Trans, 8 A, Feb p 373, (1977).

F. P. Bundy: Reprints of first international conference
on materials, ASTM, Philadelphia, Pa, Feb 3-6, (1964).

C. G. Rhodes and N. E. Paton, Metallurgical TRANS,
Volume 10 A, P 209.

Banerjee and Arunachalam. Acta Metallurgica 22, p 1685,
(1981). '

D. Banerjee and J. C. Williams, Scripta Metallurgica,
12. pp 1125-1128. (1983).

D. G. Westlake, and W. R. Gray, App. Physics Letters,
2. 3 (1955).

4218

174

I