Directional eutectoid decomposition in multicomponent
systems : isothermal exploration lambda-R relations
crystallography
Citation for published version (APA):
Wolff, L. R. (1977). Directional eutectoid decomposition in multicomponent systems : isothermal exploration
lambda-R relations crystallography. Technische Hogeschool Eindhoven. https://doi.org/10.6100/IR36919
DOI:
10.6100/IR36919
Document status and date:
Published: 01/01/1977
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be
important differences between the submitted version and the official published version of record. People
interested in the research are advised to contact the author for the final version of the publication, or visit the
DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page
numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne Take down policy
If you believe that this document breaches copyright please contact us at: openaccess@tue.nl
providing details and we will investigate your claim.
Cover photograph: quenched in transformation front of a directionally decomposed Cu-11.8 wt% Al eutectoid.
vormgeving: i-stevens
DIRECTIONAL EUTECTOID DECOMPOSITION IN MUL TICOMPONENT SYSTEMS
lsothermal exploration À-R relations crystallography
DIRECllONAL EUTECTOID DECOMPOSITION
IN MUL TICOMPONENT SVSTEMS
isothermal exploration .\-R relations crystallography
PROEFSCHRIFT
ter verkrijging van de graad van doctor in de technische wetenschappen aan de Technische Hogeschool Eindhoven, op gezag van de rector magnificus, prof. dr. P. van der Leeden, voor een commissie aangewezen door het college van dekanen in het openbaar te verdedigen op dinsdag 1 februari 1977 te 16.00 uur
door
Ladewijk Reinaerd Wolft geboren te Utrecht.
DIT PROEFSCHRIFT IS GOEDGEKEURD DOOR DE PROMOTOREN Prof. Dr. G.O. Rieck
en
Aan Edmée Michael Margot Louisè
CONTENTS
CHAPTER 1: INTRODUCTION
1.1. Unidirectional growth of in situ composites. 1.2. Multivariancy.
1.3. The scope of the present investigations.
CHAPTER 11: ISOTHERMAL EXPLORATION
Page 11 12 12
2 .1. lntroductory Notes. 14
2.2. The Effective Ouaternary Solvus Plane. 16
2.3. Experimental Procedure for the location of the E.O.S.-planes in the
Cu-AI-Ni-ln-system. 21
2.4. Eutectoid Duplex Cellular and Duplex Dendritic Growth. 23
2.5. Uncoupled Growth. 25
2.6. The effect of alloying agents on the eutectoid decomposition temperature. 25 2. 7. Determination of the third phase precipitating in some of the ternary and quaternary
eutectoid alloys. 26
2.7 .1. X-ray diffraction pattem of the (Ni,Cu)AI precipitates found in two of the
E.O.S. samples. 28
2.8. Results. 30
2.8.1. The E.O.S.
cx/-y2
location and approximate decomposition temperature. 30 2.8.2. The E.O.S.cx/8
location and approximate decomposition temperature. 31 2.8.3. Occurrence of Eutectoid Duplex Cellular Growth. 322.9. Discussion and Conclusions. 33
2.9.1. Difficulties concerning the E.O.S. location. 33
2.9.2. E.D.C.G. and the "no partitioning temperature". 34
2.9.3. The (Ni,Cu)AI precipitates. 37
2.9.4. Uncoupled Growth, a limitation to directional eutectoid decomposition. 38
2.9.5. The temperature effect of alloying agents. 38
2.9.6. Conclusions. 39
CHAPTER 111: UNIDIRECTIONAL EUTECTOID DECOMPOSITION AND À-R RELATIONS 3.1. The history of unidirectional eutectoid decomposition. 41 3.2. The theories concerning eutectoid decomposition, eutectic solidification and cellular
precipitation. 43
3.2.1. The Zener theory for the formation of pearlite in the Fe-C system. 44 3.2.2. Carpay's theory tor aligned lamellar eutectoid transformation. 49
3.2.3. Comparison of the lamellar growth theories. 52
3.3. The unidirectional eutectoid decomposition techniques. 56
3.3.1. Discussion of the decomposition techniques. 60
3.4. The apparatus for the production of bars of eutectoid composition. 62
3.5. The bridgman apparatuses. 63
3.5.1. The B-bridgman apparatus. 63
3.5.2. The AB-bridgman apparatus. 65
3.6. Unidirectional decomposition of the Cu-11.8 wt% Al eutectoid. 68 3.7. Unidirectional decomposition of some E.Q.S. alloys in the Cu-Al-Ni-In system. 72 3.7.1. Directional decomposition of the Cu-AI-(Ni) alloys. 73
3.8. 3.9. 3.10.
3.7.2. Directional decomposition of the Cu-AI-(In) alloys.
3.7.3. Directional decomposition of a quaternary Cu-AI(Ni)-(ln) alloy. Directional decomposition of a Cu-ln-(Ni) alloy displaying E.D.C.G. The transformation front.
Discussion and conclusions.
3.1 0.1. The Cu-11.8 wt% Al eutectoid.
3.1 0.2. Comparison of the À-R relations in Cu-Al and E.O.S. céy2 alloys. 3.1 0.3. Minimum interlamellar spacing.
3.1 0.4. Partitioning and E.D.C.G. in directional decomposition of eutectoids. 3.1 0.5. Same remarks on the À-R relations.
3.10.6. Problems concerning the experimental determination of the À-R relation in eutectoid decomposition processes.
3.1 0. 7. Justification of the u se of an interval switch in order to obtain low
3.10.8. 3.10.9.
growth-rates. Conclusions.
Note added in proof: The baron dope.
74 76 78 82 86 86 91 94 95 96 98 99 100 101
CHAPTER IV: CRYSTALLOGRAPHIC RELATIONS IN DIRECTIONALL V DECOMPOSED EUTECTOIDS
4.1. I ntroduction. 103
4.2. Texture goniometry. 103
4.2.1. S.B.R. technique. 104
4.2.2. The texture goniometrie method. 105
4.2.2.1. The graphical procedure for the completion of pole figures. 015 4.2.2.2. Some problems concerning sample perfection and resolution in Bragg angles. 107
4.3; The Cu-Al eutectoid. 107
4.3.1. Experimental. 108
4.3.2. Results. 109
4.3.3. Discussion. 112
4.3.3.1. The twinning of the a-Cu phase. 112
4.3.3.2. The interface plane. 113
4.3.4. Crystallographic examination of the Kurdjumow-Sachs relation between
a-Cu and 'Y2 -Cu9AI4 phase in the Cu-Al eutectoid. 114 4.3.5. The high temperature diffraction experiment. 118 4.3.6. Dilatometric measurement of the density change during the eutectoid
reaction in the Cu-11.8 wt% Al eutectoid. 120
4.3. 7. The ditfusion coup Ie experiment. 121
4.3.8. lnfluence of the decomposition rate. 124
4.4. The crystallography of the Ni and/or In modified Cu-Al eutectoid. 127
4.4.1. The Cu-11.8 wt% Al-0.5 wt% eutectoid. 127
4.4.2. The Cu-10.95 wt% Al-2 wt% In eutectoid and the Cu-9.6 wt% Al-6.0 wt%
In eutectoid. 128
4.4.3. The Cu-10.4 wt% Al-4.0% ln-2.0% Ni eutectoid. 130
4.5. The Cu-31.5 wt% In eutectoid. 130
4.5.1. The 8-phase in the Cu-In system. 130
4.5.2. The crystallography of the Cu-31.5 wt% In eutectoid. 132
4.6. The Cu-31.9 wt% ln-4.0% Ni eutectoid. 136
4.7. 4.8. 4.9. 4.10. 4.11.
The Ni-70.6 wt% In eutectoid. The Fe-43.0 wt% Al eutectoid. The Co-13.7 wt% Si eutectoid.
The Co-40.2 wt% Ni-13.7 wt% Si eutectoid. Discussion and Conclusions.
SUMMARY Rf:SUMt: SAMENVATTING REFERENCES LIST OF SYMBOLS DANKBETUIGING LEVENSBERICHT
LIST OF TABLES AND FIGURES
137 141 146 149 155 163 165 167 169 173 177 177 baak flap
I. INTRODUCTION
1.1. UNIDIRECTIONAL GROWTH OF IN SITU COMPOSITES
Composite materialscan be defined as heterogeneaus materials consisting of two or more solid phases, one of these phases having at least one dimension in the range of 10-6- 10-2 cm. The interest in composite materials arises tromthefact that their properties may differ substantially from those of the composing phases. This is especially the case when these phases are oriented in a periadie and/or anisotropic way. Composite materials e.g. can offer opportunities as strengthened materials for construction and further advancement of supersonic aeroplanes, space vehicles, high pressure tanks etc. which require mechanica! properties which cannot be met by existing homogeneaus materials.
A special class in this conneetion is formed by composites composed of a ductile matrix (e.g. metal or plastic) in which is embedded a large quantity of whiskers, fibres or coated filaments of a strong but brittie material I ike SiC, AI203, metalborides, boron, graphite or glass-fibers. A whisker or fibre-lenght of 1Q-5- 1Q-2cm can be considered typical. lf e.g. whiskers with undamaged surfaces are embedded parallel toeach other in an affixing, non-agressive ductile material, then this artificial composite wil I stand up to heavy loads when loaded in a direction parallel to that of the whiskers (Broutman et al. 1967 (1.)). A large variety of strong whiskers can be grown e.g. from the gas phase (Wagner 1970 (2.)). The sorting, aligning and embedding of the whiskers without darnaging their surface, however, poses numerous problems. This is one of the reasans why it can be advantageous to grow the whiskers in situ in the matrix in such a way that they are aligned at the same time. Th is can be accomplished in a one step process e.g. by eutectic reactions like Straumanis and Braks statedalready in 1935 (3.). Kraftand Albright in 1961 (4.) and Salkind and Lemkey in 1967 (5.). Such a process has the extra advantage of the whiskers and matrix being more or lessin chemica! equilibrium with each other, which promotes the perfection of the whisker surface even at higher temperatures.
Composite materials, however, arejustas promising in other fields than those using their mechanica! strenght. At the moment most metallic permanent magnets are made of Alnico-alloys which are decomposed in the solid state. This is done under specific cooling conditions in an artificial magnetic field (de Vos 1966 (6.)). A representative example in electranies is the fibrous composite material which is made by unidirectional solidification of a homogeneaus eutectic melt of the quasibinary NiSb-lnSb system. The resulting composite consistsof parallel metallic NiSb fibres (lenght > 5.10-3 cm,
diameter~ 1Q-4 cm) in a matrix of semiconducting lnSb. The positive magnetoresistance is very anisotropic and unusually strong when the direction of the metallic fibres, the direction of the electrical current and that of the applied magnetic field are perpendicular toeach other (Weiss 1966 (7.)). Applications of this material include devices for measuring fieldsin cryogenic magnets and contactless variabie resistors. Galasso 1967 (8.) has given a survey of non-mechanica! applications of unidirectional solidified eutectics.
In the past two decades the interest in in situ composites has steadily increased, which is reflected in the number of surveys covering this field that have been published (Tiller 1958 (9.). Winegard 1961 ( 1 0.). Chadwick 1963 ( 11.). Kr aft 1966 ( 12.), Kerr and Winegard 1966 (13.). Hunt 1968 (14.), Durand 1969 (15.), Albers 1969 (16.). Hogan c.s. 1970 (17.). Livingston 1970 (18.) and Kurz and Sahm Î19.)).
1.2. MULTIVARIANCY
Composites can be grown in situ by directional solidification of eutectics. Such composites can retain much of their strenght up toa tew tens of Kelvins below their melting point, because the coexisting phases are in chemica! equilibrium. More recently it was demonstrated that composite materialscan also be obtained in situ by direction al transformation in the solid state (Carpay 1970, (20.). Balling and Richman 1970 (21.). Livingston 1970 (22.)). One of the limitations of composites grown in situ as far as applications are concerned -is the difficulty of adapting them to specific needs. Th -is d-isadvantage can be overcome toa great ex tent by making use of the possibilities afforded by multivariancy. Th is has been demonstrated tor a number of multivariant eutectics which have been directionally
solidified such as Ni-Co-Cr-W-AI-TaC (Bibring et al. 1972 (23.)). Fe-Cr-Ni-C (Thompson and Lemkey 1970 (24.)). Fe-Co-Cr-C (Wolft 1972 (25.)). etc. Here an invariant eutectic is rendered multivariant by the addition of new components, while the resulting composites remain tree trom primary crystals such as dendrites, which could have a detrimental effect on their strenght. At the same time, addition of new componentscan result in increased oxidation resistance and improved high-temperature strenght. There is also an effect on the volume fractions of the constituing phases which may cause a change in morphology (lamellae to needie transition). In this way a composite grown in situ can be adapted toa specific use.
1.3. THE SCOPE OF THE PRESENT INVESTIGATIONS
Although the subject of directional solidification of multivariant eutectics is still in needof much research (Same valuable work has been done by v.d. Boomgaard 1973 (26.) and 1975 (27 .). the investigations reported in th is thesis concern a new cl a ss of multivariant systems, viz. multivariant eutectoids. lt has been mentioned above that directional transformation in the solid state is a means of obtaining a composite grown in situ. One way of doing this is by directional eutectoid decomposition. The systems in which such a directional eutectoid decomposition has been achieved include Cu-Al (Carpay 1970 (20.); Livingston 1970 (22.)). Cu-In, Ni-In, Co-Si (Carpay 1973 (28.)), etc. Application of such directionally decomposed binary eutectoids is limited in the same way as that of binary eutectics by their lack of adaptability. The first object of the present investigations was to find out whether this drawback can be overcome, as in the case of eutectics, by making u se of multivariance. For this purpose the system Cu-Al-Ni-In wasselectedas an example. In chapter 11 an account is given of the isothermal exploration of this quaternary system. The purpose of this
exploration was to locate alloy compositions which upon directional eutectoid decomposition might produce aligned composite materials.
Chapter lil concerns the directional decomposition of multivariant eutectoids. One of the main topics is the relation between decomposition rate R and interlamellar spacing À. As this subject is al ready cantroversial in binary eutectoids (see ~-9-publications of Carpay and v.d. Boomgaard 1971 (29.), Carpay 1972 (30.), Cheetham and Ridley 1973 (31.),
Livingston 1974 (32.), Meilor and Chadwick 1974 (33.)), the À- R relationship in the binary Cu-Al eutectoid was investigated first. Then the influence of Ni and/or In additions on this relation is paid attention to.
Chapter IV covers the subject of the crystallographic relations insome binary ternary and quaternary eutectoid systems which have been directionally decomposed. The motivation for this study is that crystalline perfection may be important when application of these directionally decomposed eutectoid materialsin the field of electronic devicesis considered.
11. ISOTHERMAL EXPLORATION
2.1. INTRODUCTORY NOTES
As the eutectoid compositions in polynary systems are virtually unknown, these composit· i ons have to be determined. With eutectoid compositions we mean those compositions, which upon directional eutectoid decomposition may give aligned two phase morphology without primary crystals (denJrites) of either of the two low temperature phases. Th is chapter deals with the effort todetermine these compositions in the Cu-Al-Ni-In system by isothermal transformation experiments. The reason why we selected the Cu-Al-Ni-In system as a model system is threefold:
1. Th is system contains three binary eutectoids (Fig. 2.1) which have been directionally decomposed by other workers: the Cu-Al eutectoid (Carpay 1970 (20.). Livingston 1970 (22.). Cheetham and Ridley 1973 (30.)); the Cu-In eutectoid (Mellor and Chadwick 1974 (32.)); the Ni-In eutectoid (Carpay 1971 (28.)).
2. The composing elements seemed easy to handle and were available in the required purity. 3. The eutectoid decomposition temperatures are not too high (Cu-Al: 565°C; Cu-In: 574°C
and Ni-In: 7700C).
W[IGHT PER CENT ALUMINUM
10 15 20 25 30 40 50 60 70 80 90 1200 t!OO 1000 900 ~ 800 ~ "' => ~ 700 "' ~
"-"'
!' 600 500 400 300~~
1048° 110l7°r-: ~ p a.c,1/
or(Cu} 19.6 ~· 19.4) ii
~70-400°r--I h I Ir~
,"
~Tt~
1\
11
~·-\
1\
r,I~
624°~
5~f0 ~.----
--I 563° -a {,;;• -~. 8 200 0 Cu 10 20 30 40 50 60 70 ATO~IC PER CENT ALUMINUM
a
--.L.I..21
-...'
L 5•e• .:L!,l 82.7 167.01 80 90 97.5 194.~1'
100 A\ r !AllW[IGHJ PER CUT INDIUM 10 20 lO 40 so 60 70 80 90 '10 0 700 ~"""- m'
or\
6~0l1
~.~
PrF
630'?
'
0 '. \ 600 ,.~. 110 100 I 1\ ~,...
0 20.,5 I\
3 I ~so I ot • a I 0..
0\
t ~ :( I soo IS 20 2S :10 lS 800 -,,~o7 Ho•\rZr
"'
,--
--otiCu) \r---_
I ~ 700 51'4°..
.
~' (18.21 I I I-
~"'
~soo 400 lOO 200 100 0 0 CoI
1.212.1SI•
20 ~700 500 I , I I I I 10°,,,
I!t-•
,,
lr'
,5,.,,,
'"
'"
'"
lO tO lO tO 70 s. zuo•rc PU cE•t ouru•
b
~,L_ _ _ ~s~o---~6co===='~=·==~7o •ina t ' f . l n-c
10'\
\
\
I I ~·· 100 I•Fig. 2.1. Phase.diagrams of the systems: a) Cu-Al, b) Cu-In as given by Hansen 1958 (34.) and c) Ni-In as given by Bestand Gödecke 1969 {35.).
Before coming to the experimental part of our isothermal exploration of the Cu-Al-Ni-In system, the concept of Effective Ouaternary Solvus (E.O.S.)-planes will be introduced.
2.2.
THE EFFECTIVE OUATERNARY SOLVUS-PLANEIn order to make the concept of the E.O.S.-plane comprehensible we shall first look at the binary equivalent in a liquid-solid transition.
Consider the phase diagram of the system A-B (Fig. 2.2.1.). lf we have a melt of composition x 1 and we cool it to the temperature T 1• solidification will start. At this temperature, however, the composition of the solid phase in equilibrium with a liquid phase of
composition x 1 wil I be x2. lf we want to produce a unidirectionally solidified bar of overall composition x 1, we are hindered by the tact that the composition of the liquid phase is different from that of the solid phase with which it is in equilibrium. This will cause segregation and the first solidified part of the bar will have composition x2 while the last part to solidify will be of e.g. composition x4 - ordepending on the conditions- can even be of eutectic composition, xE. Of course the overall composition must be the original composition of the melt.
T, T I a' I I I < I ' I I I L 'E B
·-Fig. 2.2.1 Phase diagram of an imaginary svstem A-8 (T
=
temperature, x= atomfraction, T mA=
melting point of pure A).The only chance we have to produce a reasonable lenght of bar of constant composition x 1,
is that we reach a steady stateduringa processof unidirectional solidification (Fig. 2.2.2.). This can be accomplished using a floating crucible technique, also called the double crucible technique (Goorissen 1960 (36.)), or the edge-defined film·fèd crystal-growth technique (Chalmers et al. 1972 (37 .) ). Although the composition of the melt (x3) and solid (x 1) at the solid-liquid interface are not the same, a steady state can be reached if the composition of the melt fed into the transition zone (T.Z.) is the same as the composition of the solid substracted from the transition zo~e (x1). So far we have only discussed the simple case of the liquid·solid transition in a binary system , but the situation of a eutectic transition in a ternary system
is in many ways analogous to the situation we discussed. Different of course is the fact that
components in stead of two.
i
T·
-Fig. 2.2.2 Concentration profile during steady state directional solidification of a single solid phase in a bicomponent system.
However, the number of the degrees of treedom according to Gibbs' phase rule is the same. Here too we have the situation that the composition of the melt at the solid-liquid interface is in general different trom the gross composition of the two solid phases with which it is in equilibrium.
lf we want to obtain a bar of aligned eutectic material of constant composition, here too we have to attain a steady state in which the composition of the melt fed into the transition zone is the sa me as the gross composition of the sol id phases substracted trom this transition zone. This can again be accomplished using techniques like the double crucible technique or the film-fed edge-defined crystal-growth technique. There is however an additional problem: in general we do not know thesetof gross compositions of the low temperature phases in a ternary system which upon directional solidification may produce a material with aligned eutectic morphology without primary crystals (dendrites) of one the two solid phases. This can best be illustrated with a phase diagram of an imaginary ternary system (Fig. 2.3A.) in which two of the components (A and B) are completely miscible, while they both form
i
Ta
ceutectics with the third component (C). In such a system one might find
a
mono-variant eutectic range starting at E 1 (the composition of the A-C-eutectic) and continuing till E2 (the composition of the B-C eutectic).B
••
'•
l
b
c
d
Fig. 2.3 An imaginary system A·B·C with a monovariant eutectic. A) T-x diagram (T
=
temperature, x=
atomfraction). 8) composition triangle.C) directional solidification of a melt of E. T.S.-composition: initia/ period. D) directional solidification of a melt of E. T.S.-composition: steady state. In general the composition of the melt and that of the solidified monovariant eutectic with which it is in equilibrium will not be the same. This is illustrated bv the fact that the lower border of the liquidus vallev Ei E2 when projected on the composition plane ABC -does not coincide with thesetof gross compositions of the solid monovariant eutectic (dotted line El - E2). Th is set of gross compositions of the sol id monovariant eutectic is called Effective Ternarv Solidus (E.T.S.) bv v.d. Boomgaard 1973 (26.). 1975 (27.) who has covered this subject thoroughlv with all its possible variants. lf we want to produce a
bar of eutectic morphologv and constant composition through a processof unidirectional
solidification, we wil I have to know the position of the E.T.S., because onlv with compositions on this E.T.S. we can reach a steadv state in which the composition of the melt fed into the transition zone is the same as that of the solid eutectic substracted from it. Th is can be illustrated with Fig. 2.3.B, C and D. Fig. 2.3.B is the composition plane of the svstem A·B·C. In this composition plane we see two solid phases: the o:- and the
-y-phase. Furthermore we see the projectionsof the lower boundary of the liquidus vallev (dotted line E1 - E2) and the E.T.S. (solid line E1- E2).
Now let us assume we want to produce a bar of eutectic composition x 1· The steady state directional solidification will then be attained in the following wav. We lower a bar of molten allov (composition x 1) trough a temperature gradient. At a certain temperature the bottorn part of the bar wil I reach the liquidus plane and the o:-phase of composition x3
will nucleate. As we lower the bardendrites of composition x3 will grow into the liquid and alter its composition till again at the bottorn of the bar it will reach composition x2 (The composition trajectorv of the liquid at the liquid/solid interface is given in Fig. 2.3.B by the thick solid line with arrows x1 ~ x2 ~ x5). When the liquid reaches composition x2, at the bottorn of the liquidus vallev eutectic o:
+
'Y of gross composition x4 wil I nucleate and the dendrites of composition x3 wilt stop growing (Fig. 2.3.C). As the bar is lowered, theeutectic wilt grow out and the liquid at the solidification front will be changed in
composition from x2 tilt x5 bv the solid eutectic that is being form.ed.The solid eutectic in turn changes from gross composition x4 till gross composition x 1. At the same time the transition zone is built up and will change trom its initia! composition TZ 1 when the first
eutectic was formed (Fig. 2.3.C) until it reaches its steady composition: TZs.s.
(Fig. 2.3.D). Now we have a situation in which the input: liquid of composition x 1 which is "fed" into the transition zone, is equal to the output: eutectic of composition x 1 which solidifies at the solidification·front. Thus we have attained a steady state processin which aligned eutectic of composition x1 is produced. This process will continue tillallof the liquid but the T.Z. is consumed. Th en the steady state is disrupted and the eutectic wil I change in composition wh ile the transition zone solidifies.
Having introduced the concept of the E.T.S. we can now direct our attention towards the analogous case of the eutectoid decomposition in a ternary system. Apart from the fact that a eutectoid reaction is a slower solid state process, in which a high temperature phase decomposes into two low temperature phases, the situation is the same as in a eutectic reaction.
lf we consider a eutectoid decomposition in a ternary system it is our hypothesis, that in general we wil I also find the composition of the high temperature phase to be different from the gross composition of the low temperature phases with which it is in equilibrium. None the less there should be a range of compositions which upon directional eutectoid decomposition give aligned eutectoid morphology without primary crystals (dendrites) of either of the low temperature phases. Justas in the case of eutectic solidification discussed above, we would have to create a steady state in which the composition of the high temperature phase fed into the transition zone is the same as the gross composition of the low temperature phases substracted trom it.
This implicates that we have to locate the Effective Ternary Solvus in such a ternary system. Th is Effective Ternary Solvus t:an bedefinedas thesetof gross compositions of the low temperature phases in equilibrium with thesetof high temperature phase compositions that together make up the lower border of the solvus valley in the phase diagram. Th is can be
Fig. 2.5 Argon-Are melting apparatus for the production of samples for isother-mal transformation experimen ts.
Fig. 2.6 R.F. turnace for the production of samples for isothermal transformation experiments.
illustrated by Fig. 2.4. in which we see an imaginary phase diagram with components D, F and G. The (3 phase decomposes eutectoidically into the low temperature phases o: and 'Y-By addition of component F, the invariant eutectoid in the binary system D-G has become a monovariant eutectoid. In the composition triangle 0-G-F we see the line E'P' which is the projection of E.P.: the lower border of the homogeneity region of the high temperature phase (3. The line E'O' is the Effective Ternary Solvus. Note that each gross composition of the low temperature phases on the E.T.S. can be connected by a "pseudo"-tieline- that is a tangent to the curve E'P'- to the high temperature phase composition with which it is in equilibrium.
So far we have seen how an invariant eutectoid in a binary system becomes a monovariant eutectoid in a ternary system. lf we add a fourth component to this system we obtain a bivariant eutectoid. lnstead of the E.T.S.-Iine wethen will find an Effective Quaternary Solvus (E.O.S.)-plane. This E.O.S.-plane is thesetof gross compositions of the low temperature phases that are in equilibrium with the various compositions of the correspon-ding high temperature phase.
Thus having introduced the concept of the Effective Ternary Solvus-line and the Ouaternary Effective Solvus-plane, we wil I now proceed with a description of our experimental
procedure for locating the E.O.S.-planes in the model system: Cu-Al-Ni-In.
2.3. EXPERIMENT AL PROCEDURE FOR THE LOCATION OF THE E.O.S.-PLANES
IN THE SYSTEM Cu-Al-Ni-In.
In order to locate two of the E.O.S.-planes in the Cu-Al-Ni-In system we adopted the following procedure. Starting with the known eutectoid compositions in the binary systems Cu-Al: Cu 11.8 wt% Al (Fig. 2.1 a) and Cu-In: Cu 31.5 wt% In (Fig. 2.1 b). we prepared 10 gram samples with various additions of Ni and/or In, or Ni and/or Al respectively. We used starting materials with purities better than 99.99%. These samples were melted using two techniques. The first one consistedof argon-are melting on a watercocled Cu plate
Fig. 2.7 Cu-In-Ni al/oys: Jeft, Cu-31 wt% /n-5 wt% Ni: excess Cu (primary o:-Cu crystals); centre, Cu-33 wt% /n-5 wt% Ni: Cu deficiency (primary o-Cu7ln3 crystals); right,
Cu-32 wt% /n-5 wt% Ni: eutectoid morphology.
(repeated at least three times in order to promote homogenisation). The apparatus concerned is depicted in Fig. 2.5.
The second technique used, was melting of the sample in an R.F. turnace in an alumina crucible in a flowing argon atmosphere (Fig. 2.6.).
Here too the procedure was repeated at least three times in order to promote homogenisation. The so prepared samples were then sealed in a silica capsule under vacuum. Subsequently they were subjected toa 24 hour homogenisation annealing at 650° · 700°C. Then they were isothermally transformed at about 30°C below their equilibrium eutectoid
decomposition temperatures (530°- 600°C). After this thermal treatment the samples were quenched in water.
The isothermally transformed samples were evaluated metallographically in termsof excess or shortage of e.g. Cu or eutectoid morphology (Fig. 2.7.).
The etchant used was 15 vol.% conc. HN03, 25 vol.% glacial acetic acid and 60 vol.% acetone.
New alloys were then weighed in according to the metallographic results. In this way we have located two E.O.S.-planes in the quaternary composition tetrahedron of the Cu-Al-Ni-In- system (Fig. 2.8.).
Al wl" NI
..
..
~ ---·--- -Cu In..
E.O,S. a/0 plane •oFig. 2.8 Composition tetrahedron of the Cu-Al-Ni-In system with two of the E. O.S.-p/anes.
First we explored three ternary facesof the composition tetrahedron: the Cu-AI-Ni-system,
the Cu-AI-In-system and the Cu-ln-Ni-system. Wethen explored planes of constant In percentages, parallel to the Cu-Al-Ni plane of the composition tetrahedron. The sections ex plored had In percentages of 2, 4, 6 and 8 wt% In. The numerical results of these
explorations will be given insection 2.8. We also explored the extension of the E.O.S. plane at the Cu-In-Ni side in the Al-direction. This extension however proved to be very restricted and probably leads to precipitation of a third phase.
The two E.O.S.-planes we have thus located were the E.O.S.-a/-y2 concerning the eutectoid reaction:
{3-Cu3AI + a-Cu
+
-y2-CugAI 4at the Cu-Al side of the composition tetrahedron. The other E.O.S.-plane is the E.O.S.-wo concerning the eutectoid reaction:
{3-Cu4 1n + a-Cu
+
8 -Cu7 1n3 *at the Cu-In side of the composition tetrahedron.
2.4. EUTECTOID DUPLEX CELLULAR AND DUPLEX DENDRITIC GROWTH We al ready mentioned, that, at the transformation front, the gross composition of the low temperature phases and that of the high temperature phase from which they are formed, are in general not the same in a multicomponent eutectoid. lf these ditterences in
composition become large enough, this may result in a curved interface at the transformation front, the composition of the high temperature phase is altered and a different composition implicates a different transformation-rate at the same transformation-temperature. Now if the transformation-rate of the altered high temperature phase is lower than the original one, a curved interface may develop (Fig. 2.9.). Because of the resulting morphology, this phenomenon is called Eutectoid Duplex Cellular Growth (EDCG) (Duplex stands for two-phase).
Th is change of the composition of the high temperature phase upon transformation may effect a similar morphology in unidirectionally transformed multicon'lponent eutectoids. Here a different composition implicates a different transformation temperature at the same transformation ra te.
Fig. 2.9 Eutectoid Duplex Cel/u/ar Growth (E.D.C.G.) (Schematic)
* In the literature the formula of the o-phase is given as Cugln4 (34.) but we preter the tormul a given above as will be explained in chapter 4 .5.
Again, if the new transformation temperature is lower than the original one, a curved inter-face may develop at the transformation front. Here, however, this effect may be reduced by employing a steep temperature gradient during unidirectional transformation.
So far we explained when E .D.C.G. was to be expected in a multicomponent system. lndeed we found E.D.C.G. in several of our isothermally transformed alloys (Fig. 2.10). The occurrence of E.D.C.G. seemed to depend not only on the alloy composition but also on the isothermal transformation temperature (an interesting point that will be pursued further in sectien 2.9.2.).
Insome alloys we even found Eutectoid Duplex Dendritic Growth (Fig. 2.11.). This morphology could be the result of a more pronounced case of a curved interface at the transformation front. On theether hand it might also betheresult of facetted growth. This mode of growth is sametimes found in eutectics (Cense 1971 (38.)) where the solidification front may be facetted. So far, however, we have never witnessed a facetted transformation front in any of our eutectoid samples.
Fig. 2.10 E.D. G.G. in an al/oy of compo
-sition Cu-33 wt% ln-6 wt% Ni,
slowly caoled til/ 5300C, then quenched.
Fig. 2.11 Eutectoid Duplex Dendritic Growth (E.D. G.G.) in an alloy of
composition Cu-33 wt% /n-6 wt%
Ni, slowly caoled til/ 5300C, then
2.5. UNCOUPLED GROWTH
In the course of our isothermal explorations of the Cu-Al-Ni-In system concerning the E.O.S. locations, we sometimes encountered a phenomenon we have called uncoupled growth. This occurred in samples decomposed at temperatures just below their equilibrium transformation-temperature (Fig. 2.12.). Wethink this uncoupled growth occurs when the ditfusion rate of the elements needed for the growth of a nucleus is greater than the rate at which these elements are builtinto the growing crystal. ·
Fig. 2.12 Uncoupled growth in an a/loy of composition Cu-11.8 wt% Al, partly transformed after 24 hours at 55JOC, then transformed martensitically upon quenching.
In other words uncoupled growth occurs when there is no ditfusion limitation to the growth of a crystal of the low temperature phases. As the growth rate falls with rising temperature while the ditfusion rate rises, uncoupled growth is to be expected for all eutectoid reactions just below their equilibrium transformation-temperature. As it is very hard to judge whether a sample which has been eutectoidically decomposed by uncoupled growth, is hypo-or hyper-eutectoid or eutectoid in composition, we have tried to avoid these conditions whenever possible, at least when trying to locate the E.O.S.-planes in our model system.
2.6. THE EFFECT OF ALLOVING AGENTS ON THE EUTECTOID DECOMPOSITION
TEMPERA TU RE
By new series of isothermal transformation experiments carried out at various temperatures (about 10°C apart). the effect of alloying agents upon the decomposition temperatures of a range of E.O.S.-compositions vitas studied. As D.T.A. proved to be impractical to this end, because of the low transformation rates and the volatility of In, we adopted the approach of bringing the samples* above their eutectoid transformation temperature and then · cooling them toa given temperature at which they were kept for 24 hours. After this *We used the same 10 gram samples in evacuated silica capsules that we re used lor the E.O.S.-Iocation.
treatment they were water-quenched and their microstructure was analysed with respect to the presence or absence of eutectoid mate rial. Thus we establ ished, whether the chosen temperature was below or above the decomposition temperature of that particular alloy. In this way, it was found that the modification of the Cu-Al eutectoid with Ni and/or In results in raising the eutectoid decomposition temperature trom 565°C up to 635°C. Modification of the Cu-In eutectoid by Ni additions also raises the eutectoid transformation-temperature from 574°C up to 630°C. More detailed information will be given insection 2.8.1. and 2.8.2.
2.7. DETERMINATION OF THE THIRD PHASE PRECIPITATING INSOME OF THE TERNARY AND OUATERNARY EUTECTOID ALLOYS
When trying to locate the E.O.S.-planes in the Cu-Al-Ni-In system we found that in the Cu-Al-Ni system alloying of the Cu-Al eutectoid with more than 1 wt% Ni resulted in very fine precipitates which apparently had little effect on the eutectoid decomposition and the morphology resulting from it. These precipitates increased in number and in size with increasing Ni content. They seem to be arranged in "chains" Fig. 2.13.
Fig. 2.13 "Chains" of Ni A/ precipitates in an alloy of composition Cu-13 wt% A/-5 wt% Ni, isothermally transformed at 5500C.
In reality they form envelopes around what must have been the grains of the high te mpera-ture ,6-phase. lt is important to note that they apparently do not block the eutectoid transformation as they have no relation with the eutectoid colony boundaries what so ever. In the quaternary Cu-Al-Ni-In systems we also found such precipitates but indium seemed to retard the nucleation rate in this precipating proces. Thus the precipitates formed at
higher Ni contents we re larger but less in number than those formed in eutectoid alloys with corresponding Ni contentsnot containing In.
At the Cu-In side of the Cu-Al-Ni-In we found that eutectoid alloys not containing Al could contain up to 11 wt% Ni without causing a third phase to precipitate. However, addition of
only 0,5 wt% Al to the Cu-Ni-ln-alloys caused precipitation of a third phase, the precipitates being small in number but large in size (Fig. 2.14.).
Fig. 2.14 NiAI precipitate in an alloy of composition Cu-32.5 wt% ln-2 wt% Ni-1 wt% Al,
isothermally transformed at 5700C.
Fig. 2.15 NiAI precipitate chain on a grain-boundary of martensitically
transformed (3 in an alloy of com-position Cu-13 wt% A/-5 wt% Ni
(quenched from 6000C).
Fig. 2.16 NiAI precipitates in a {3-matrix of an alloy of composition
Cu-31.25 wt% /n-2 wt% Ni-0.5 wt%
Al, quenched from above the eutectoid transformation tempe-rature.
Both the precipitates at the Cu-Al si de and those at the Cu-In si de of the Cu-AI-Ni-1 n composition tetrahedron nucleated in the high temperature {3 phases. Th is can be illustrated by Fig. 2.15 and 2.16 showing specimens that were quenched from above the transformation-temperature.
We expected these precipitates to be a NiAI-related phase and in order to confirm this we analysed them by means of a mieroprobe S.E.M. IIA apparatus of Associated Electrical Industries, Manchester, England. The method used tor the calculation of the precipitate composition was the empirica! method conceived by Ziebold and Ogilvie 1964 (39.). The accuracy of the method is estimated to be 1%. The two samples in which we analysed the precipitate composition were:
Sample No. 234 containing 78.43 wt% Cu 12.57 wt% Al 7.00 wt% Ni 2.00 wt% In. Sample No. 606 containing 66.00 wt% Cu 2.00 wt% Ni 1.00 wt% Al 31.00 wt% In. The precipitate composition of sample No. 234 was found to be:
35.5 at% Ni= 44.5 wt%Ni 40.8 at%AI = 23.5 wt%AI 23.7 at%Cu= 32.0 wt%Cu
0.0 at% In= 0.0 wt%1n
The precipitate composition of sample No. 606 was found to be:
48.1 at%Ni = 61.5 wt%Ni 41.7 at%AI = 24.4 wt%AI 10.2 at%Cu= 14,1 wt%Cu
0.0 at% In= 0.0 wt%1n
After the composition of the precipitates had thus been analysed samples of the same composition as the precipitates were made in the R.F. turnace (Fig. 2.6.) under a flowing Ar atmosphere. After two hours annealing at 410°C in an evacuated silica capsule, an X-ray diffractogram was made of both samples, using CoKa radiation. The samples were both found to have a CsCI-type structures like NiAI. In tact the diffraction pattern except tor a smal I ditterenee in parameter was much like that of NiAI. The X-ray diffraction data wil I be given in the next paragraph.
2.7.1. X-ray diffraction pattem of the (Ni, Cu)AI precipitates found intwoof the E.O.S.-samples.
In sample No. 234 which had an alloy composition of:
78,43 wt% Cu, 12.57 wt% Al, 7.00 wt% Ni, 2.00 wt% In, we found (by mieroprobe analysis) the precipitate composition to be:
44.5 wt% Ni, 23.5 wt% Al, 32.0 wt% Cu.
Subjecting an alloy of precipitate composition to X-ray diffraction in a diffractometer we gathered the data given in Table 11.1.
In sample 606 with composition 66.00 wt% Cu, 2.01 wt% Ni, 0.99 wt% Al and 31.00 wt% Note: As the precipitates did not contain any measurable amount of Indium the composition could be
calculated in a ternary Cu-Al-Ni system, thereby reducing the number of standards required for this calculation.
In, we found the precipitate composition to be 61.5 wt% Ni, 24,4 wt% Al, 14.1 wt% Cu. The diffraction data gathered from an alloy of this precipitate composition are given in Table 11.2.
Table 11.1
Ni35.5 Cu23.7 Al40.8 dÄ 1/1 hkl
Rad.: FekO! À= 1.93728 Ä 2.893 21 100
Filter: Mn 2.044 100 110
System: Cubic CsCI·type 1.669 4 111
1.445 7 200
a= 2.8901 Ä 1.292 5 210
1.180 17 211
1.022 6 220
Table 11.2
Ni48.1 Cu10.2 Al41.7 dÄ 1/1 hkl
I
Rad.: FekO! À= 1.93728 Ä 2.883 14 100
Filter: Mn 2.037 100 110
System: Cu bic CsCI·type 1.663 5 111
1.440 16 200
a= 2.8808 Ä 1.288 7 210
1.176 72 211
1.018 78 220
In Table 11.3 the NiAI diffraction data are listed for comparison.
Table 11.3
Ni Al dÄ 1/1 hkl
Rad.: Cuka? 2.87 40 100
2.02 100 110
System: Cu bic CsCI·type 1.655 20 111
1.434 20 200
a= 2.88 Ä 1.285 10 210
1.171 70 211
Ref.: Bradley and Taylor 1.015 20 220
Proc. Roy.Soc.London 159 A, 56 (1937).
2.8. RESUL TS
In this section we will give a survey of the numerical results of our isotherm al exploration of the eutectoids in the Cu-AI-Ni-ln-system.
2.8.1. The E.O.S.
cxl-y
2 location and approximate decomposition temperatureThe E.O.S.
cxl-y
2 is a curved plane in the quaternary Cu-Al-Ni-In composition tetrahedron. The location of this plane and the relevant eutectoid decomposition temperatures wil I be given in a series of sections trough this quaternary composition tetrahedron (Fig. 2.17 -2.18). The points in these graphs correspond to compositions of samples that have been found to exhibit eutectoid morphology upon isothermal transformation.i 8 2· &lS'è Cu.AI-Nt-!n E .a s. a/v, E.T.s.o.jy, Cu-AI·In Balance.Cu 9 10 11 12 1) wt0/o A l
-Fig. 2.17- 2.18 Composition diagrams concerning the E.
a.s. -
cx/-y
2 location. The indicatedcompositions are those giving eutectoid morphology u pon isothermal transforma ti on; some estimated eutectoid temperatures are a lso given.
2.8.2. The E.O.S.
wó
location and approximate decomposition temperatureJust like the E.O.S.
w·n.
the E.O.S.o:/ó
was located by exploring sections trough the quaternary Cu-Al-Ni-In composition tetrahedron. Only in this case the extension of the E.O.S.o:/ó
plane in the Al direction proved to be rather limited. (Fig. 2.19-2.21)E.lS.a/I'J Cu.ln .Ni 610'C 2 590'C 0 574' &30'C 31 32 ll 34 35 36 37
W1•t.rn-r
E.T.S. a/6 Cu-Al-In<
,..
'i 2 --0 27 28 2!1 J) ll l2~Ion-j,
E.Q.S. a/6z
Cu .Al-Ni-In ;: ~ 1'/,~Hiol.,..,., 02!1 ' Cu JOwt'/·~-Fig. 2.19- 2.21 Composition diagrams conceming the E.O.S.-wó /ocation. The indicated compositions are those giving eutectoid morphology upon isothermal
transformation; some estimated eutectoid temperatures arealso given . .
Modification of the Cu-In eutectoid with Al -apart trom inducing the NiAI (Cu) phase to precipitate - had a degenerating effect on the eutectoid morphology (Fig. 2.22).
Fig. 2.22 Jsothermally transformed al!oy of composition Cu-30 wt% /n-1 wt% Ni-1 wt% Al: degenerated morpho/ogy + NiAI precipitates caused by Al-addition.
2.8.3. Occurrence of Eutectoid Duplex Cellular Growth
Ou ring our isothermal exploration of the Cu-Al-Ni-In system we encountered E.D.C.G. in various samples. The composition of those samples and the temperature range in which E.D.C.G. developed is given in Table 11.4. Apart from one sample of Cu-Al eutectoid modified with 9 wt% In, all samples exhibiting E.D.C.G. belong to the Ni-modified Cu-In eutectoid.
Table 11.4
Sample No. wt% Cu wt% Al wt% Ni wt% In Temperature
range °C 172 83.00 7.98 9.02 530 635 313 65.00 3.00 32.00 610* 317 64.00 4.00 32.00 590 - 61 o* 320 63.00 5.00 32.00 590 610* 327 61.00 6.00 33.00 590 610* 334 59.00 7.00 34.00 600 610*
The Cu-Ni-In samples marked with * have been isothermally transformed at 590°C, 600°C and 610°C. The temperature range listed in Table 11.4 is restricted tothese temperatures, so the actual range of temperatures in which these alloys exhibit E.D.C.G. may well extend upwards to the temperature at which uncoupled growth occurs (section 11.5). The E.D.C.G. temperature range may also extend downwards for those samples still exhibiting E.D.C.G. at 590°C. In fact we have explored the downward extent of the E.D.C.G. range for the alloy composition of sample 317. This was done in pursuit of the "No partitioning temperature"
transformation-temperature, however, the less E.D.C.G. was observed. Below 565°C no E.D.C.G. was found.
2.9. DISCUSSION AND CONCLUSIONS 2.9.1. Difficulties concerning E.O.S.Iocation
In section 2.3 we described the experimental technique used in order to locate the E.O.S.-planes in the Cu-Al-Ni-In system. Although this technique may seem rather straightforward, we must stress that much depends upon the metallographic evaluation after the temperature treatment. In order to illustrate the difficulties that may arise in this evaluation we will discuss the influence of the temperature chosen for the isothermal decomposition of a sample.
1. lf we choose this temperature too high, that is above or just below* the equilibrium decomposition temperature TE· then the sample wil I transform martensitically u pon quenching, giving the typical martensite morphology shown in Fig. 2.23. Although no eutectoid decomposition is effected, the experiment still may be useful.
Fig. 2.23 Cu-11.8 wt% Al quenched from 5700C (just above the eutectoid transformation
temperature): martensitic morphology.
Firstly we have learned trom this experiment that the decomposition temperature is lower than we expected it to be.
Secondly, if we choose the isothermal transformation-temperature ju st above or ju st below the equilibrium decomposition-temperature, then we can still gather information about the eutectoid composition. lf the composition of the sample was off-eutectoid (not on the E.O.S.) then primary crystals of one of the two low-temperature phases wil I probably have nucleated especially at the grainboundaries of the high-temperature phase (See Fig. 2.24). So after this unsuccessful experiment we can adjust transformation temperature and composition in order to locate the E.O.S.
• Martensitic transformation upon quenching may also occur if the transformation temperature is chosen just below TE as a result.af the long incubation period.
Fig. 2.24 Cu-10.7 wt% A/4 wt% In quenched from 57o0C: 'Y2 primary crystals in a
martensite matrix.
2. lf we choose the isothermal transformation-temperature ju st a few degrees below the equilibrium eutectoid decomposition temperature and the eutectoid has nucleated, we may see quite a lot of uncoupled growth in this sample. This phenomenon of uncoupled growth (see Fig. 2.12) makes it very ditticuit to judge whether or not we have succeeded in choosing the right (E.O.S.)-composition. Uncoupled growth implicates that we will see large crystals of both low-temperature phases, so it will be practically impossible to say which component - if any - is in excess. In th is case we have to repeat the experiment with the same sample choosing the transformation temperature ten degrees lower than it was the first time.
3. lf we choose the isothermal transformation temperature more than 50°C below the equilibrium eutectoid decomposition temperature, then the high temperature t3-phase will transfarm into the ordered metastable
13
1 phase before decomposingeutectoidi-cally. This processis well known in the case of the Cu-Al eutectoid. lt was- among others- studied by Asundi and West 1966 (40.). The metastable eutectoid decomposi· tion: t3 1 ~a+ 'Y2 in this case gives rise to eutectoid (lamellar) morphology only when the composition of the alloy is Cu-12 :4 wt% Al as compared to Cu-11.8 wt% Al for the normal
t3?
a+ 'Y2 eutectoid decomposition. lt will be clea
r that such a metastable eutectoid reaction must be avoided if one wants to locate the E .O.S.2.9.2. Eutectoid Duplex Cellular Growth and the No Partitioning Temperature
The phenomenon of duplex cellular growth is well known in eutectics. Garmong 1971 (41.) studiedit in the Cu-Al eutectic modified with Mg; H.A. Ouac Bao and Durand 1972 (42.) have studied this phenomenon in Pb-modified Cd-Sn eutectic. Durand-Charre 1972 (43.) studied duplex cellular growth in the Mn-Sb eutectic modified by Bi or by Sn. In our investigation, however, duplex cellular growth was found for the first time in
eutectoids. The morphology of Eutectoid Duplex Cellular Growthis similar to that of
Fig. 2.25 Eutectic Duplex Cel/u/ar Growth in MnSb -(Sb,Bi) (Durand- Charre 1972 (43.)).
In this case the difference in morphology is mainly due tothetact that the MnSb-(Sb) eutectic has a rod morphology whereas our eutectoid is a lamellar one.
The importance of Eutectoid Duplex Cellular Growth arises mainly trom our wish to evade it as it tends to disrupt the coupling of the low-temperature phases, and possibly, even the crystallographic orientation relation. In one directionally decomposed Cu-In-Ni sample we checked the effect of E.D.C.G. on the crystallography and - surprisingly - found the orientation relation between the low-temperature phases to be justas sharp as in the related binary alloy (Chapter 4.6.). However, we discovered that, tor each composition exhibiting E.D.C.G. upon isothermal decomposition, there seems to be an isothermal transformation temperature below which this phenomenon is no Jonger seen. In any case the E.D.C.G. morphology becomes less pronounced at lower isothermal transformation temperatures. We have studied the temperature dependenee of EDCG in one alloyin particular, containing 64% Cu 4% Ni 32% In. The temperature dependenee of the EDCG morphology can be
seen in Fig. 2.26. Below an isothermal transformation temperature of 565°C, no EDCG could be perceived in this alloy.
Wethink that this phenomenon may wel I be related to the "No partitioning temperature"
recently reported for the pearlite reaction in Fe-C moldified with Mn or Cr by Razik et al. 1974, 1976 (44., 45.). They agree with Cahn and Hagel 1962 (46.) that partitioning of
substitutional elementsin Fe-C during the austenite-> pearlite reaction is a thermadynamie
necessity at higher reaction temperatures if the eutectoid decomposition is to proceed, whereas at lower temperatures there is sufficient ditterenee in free energy between austenite
and pearlite to drive the reaction without the necessity tor partitioning.
Razik et al. have measured the "no partitioning temperature" in the case of Fe-C modified witheither Mn or Cr. To this end they used an elegant technique of high resolution mieroprobe analysis with very thin samples which enabled them to obtain accurate informa-tion about the composition of neighbouring lamellae of the territe and cementite phase in
a
b
c
d
Fig. 2.26 Temperature dependenee of E.D.C.G. in an alloy of composition: Cu· 32wt%/n ·
4wt%Ni,
a) partia/ly transformed at 630 °C,
b) transformed at 620 °C,
c) transformed at 590 °C,
the pearlite. They were only concerned, however, with the partitioning of the substitutional element over the low temperature phases: a-Fe and Fe3C.
We are convinced, however, that there should also be partitioning between austenite and pearlite above the no partitioning temperature. In the case of the Cu-In-Ni system, this may cause the Eutectoid Duplex Cellular morphology. The importance of this ascertainment-if it is proved to be correct- is that E.D.C.G. may be evaded by lowering the isothermal transformation temperature beneath the "no partitioning temperature". Translated into termsof directional decomposition this means, that the transporting rate should be so high as to induce a supercooling which brings the transformation front below the "no partitioning temperature". In this way it should be possible to grow aligned multicomponent eutectoids in cases where otherwise th is would be difficult or even impossible, due to the occurrence of E.D.C.G., without applying extremely high temperature gradients.
In section 2.4. we al ready stated that E.D.C.G. wil I occur only when the altering of the high temperature phase composition upon decomposition (segregation) wil I cause a lower decomposition rate at the same transformation temperature. lf, on the other hand, the change in composition effects a higher decomposition rate, no curved interface wil I develop and as a consequence neither wil I E.D.C.G. However, the composition change upon transfor
-mation wil I still effect the building up of a transition-zone, that is if the volume ditfusion trough this transition zone is tast enough to attain a steady state. lf this is not the case, we would rather speak of a "pile up" at the transformation front as it is called by workers like
Kirkaldy 1958 (47.). Popov 1959 (48.). Darken 1961 (49.) and Hillert 1969 (50.). This "pile up" situation, however, is not to be expected when the transformation rate is enhanced by the change in composition of the high temperature phase, as e.g. in the case of the Co-modified Fe-C eutectoid (Mehl and Hagel 1956 (51.)).
lf "pile up" occurs in an alloy subjected to directional eutectoid decomposition, this will probably impede any alignment in the resulting microstructure as in this case it
necessarily is the result of a processof repeated nucleation.
lf, on the other hand, our hypothesis concerning the relation between no partitioning · temperature and E.D.C.G. température dependance is correct, then there are two ways of
coping with the E.D.C.G. problem in directional eutectoid decomposition. We can either grow the eutectoid very slowly in a very steep temperature gradient, or we can grow it so fast that the undercooling wil I bring the temperature of the transformation front below the "no partitioning temperature". When using the farmer method, the steep temperature gra -dient smoothens the interface, cause the E.D.C.G. to be less pronounced. Using the latter one, the transformation front will be flat, because no partitioning will occur between the high temperature phase and the low temperature phases in the sense that the composition of the high temperature phase wil I be equal to the gross composition of the low temper
-ature phases with which it is in contact.
lt must be stressed, however, that after the transformation has been effected, the composi
-tion of the low temperature phases will not be the equilibrium composition. This means partitioning between the low-temperature phases must occur after eutectoid decomposition in order to attain the equilibrium state,
2.9.3. The (Ni,Cu)-AI precipitates
appreciable amount of Ni showed (Ni,Cu)AI-precipitates after isothermal decomposition. Apart from the unexpected tact that this "third phase" did notshow any solubility for In at all, it was important to establish that the high·temperature phase grainboundaries, decorated by these precipitates in the 1\Ji,ln modified Cu-Al eutectoid, did not block the eutectoid decomposition. This means that the eutectoid nodules were able to extend past these grain-boundaries during their growth. Thus- although the high temperature phase graingrain-boundaries are preferential nucleation sites (and nucleation in front of the transformation front may disturb the morphology of unidirectionally decomposed eutectoids) - it is not essential in the Cu-Al and modified Cu·AI eutectoids to have large (J-grains in order to get aligned directionally decomposed eutectoid.
2.9.4. Uncoupled Growth: a limitation for unidirectional eutectoid decomposition
Insection 2.5. we already discussed the origin of uncoupled growth. Insection 2.9.1. we mentioned the difficulties it poses when it occurs in isothermally decomposed samples meant to locate the E.O.S. compositions. There is, however, one aspect of uncoupled growth that we have neglected so far: uncoupled growth forms the lower I imitation of the unidirectional eutectoid decomposition rates still giving rise to aligned morphology. lf the transportation rate in an unidirectional eutectoid decomposition experiment is chosen too low, the supercoating at the transformationfront wilt be too smalt to induce lamellar growth, resulting in uncoupled growth with no alignment of the low temperature phases what so ever. Uncoupled growth is caused by the tact that, at very smalt undercool ing, ditfusion is no longer determining the growth rate. The transformation is then controlled by molecular attachment kinetics and surface energy of the growing crystals of each low-temperature phase. Th is means we can extend the range of coupled growth rates downwards by lowering the ditfusion rate. This can be achieved by modification of the eutectoid with adequate alloying agents.
2.9.5. The temperature effect of alloying agents
In section 2.8.1. and 2.8.2. we have seen that modification of the Cu-Al eutectoid with Ni and/ or In raises the eutectoid decomposition temperature. The sameeffect was found for the Ni modification of the Cu-In eutectoid. At the same time, however, the alloying agents lower the melting temperature and melting trajectory of the Cu-Al high temperature (3-phase and the Cu-In high temperature (3-phase respectively. Thus the net effect is a reduction in temperature existence range of these high temperature phases. We suspected that this would lead to a eutectoid + eutectic transition when the high temperature phase ceases to exist. However, in neither of these two eutectoids this effect was found.
Recently Livingston 1976 (52.) has effected a eutectoid + eutectic transition in the Co.Si eutectoid by modifying it with components like W, Ta and Al.
One of the consequences of the reduction of the temperature existence range of the high temperature phase brought about by modification of the eutectoid, is that unidirectional decomposition is only possible in an apparatus in which the sample is surrounded by a crucible. This is necessary because in order to achieve the required temperature gradient, the alloy will probably be molten in the hottest part of the bridgman set up. Th is may in turn cause serious segregation during solidification.
2.9.6. Conclusions
lt has been found that:
1. The Effective Ouaternary Solvus Planes of two eutectoids in the Cu-Al-Ni-In system could be located by isotherm al transformation experiments; however, care must be taken to select the right transformation temperature.
2. lf the selected isothermal transformation temperature is too high - that is just a few degrees °C below the equilibrium eutectoid decomposition temperature- then
uncoupled growth of the low temperature phases will take place, which seriously affects
the possibility of evaluating the resulting morphology in termsof excess or shortage of Cu with respect to the eutectoid composition.
3. lf the selected isothermal transformation temperature is too low (> 50°C below the equilibrium transformation temperature) then the high temperature (j phase wil I convert
to the metastable (3 1-phase which in turn will decompose eutectoidily into a and 'Y2 at the Cu-Al side of the Cu-Al-Ni-In system.
Th is metastable eutectoid although consisting of the same phases, has a different gross
composition of the low temperature phases: it does not correspond to the E.O.S. composition. This may lead to serious errors in E.O.S. location if it is not recognised. 4. Eutectoid Duplex Cellular Growth can be explained as being a result of a ditterenee in
composition of the high temperature phase and the corresponding E.O.S. composition.
5. lt is possible to evade E.D.C.G. by lowering the transformation temperature below the "no partitioning temperature". For unidirectional eutectoid decomposition this implicates that the transporting rate should exceed a certain value corresponding to the required supercool ing.
6. The occurrence of uncoupled growth will limit the minimum transporting rate during directional eutectoid decomposition still giving rise to aligned eutectoid morphology. Th is limit can, however, be lowered by modifying the eutectoid with a component which lowers the ditfusion rate in the high temperature phase.
7. In the Ni,ln-modified Cu-Al eutectoid the high temperature {j-phase grain boundaries decorated by (Ni,Cu)AI precipitates proved to be no barrier for the passage of the eutectoid transformation front. Th is means that the eutectoid colony size is not restricted by the grain boundaries of the high temperature (j-phase.
8. Modification of the Cu-Al eutectoid with Ni and/or In raises the eutectoid temperature and at the sametime lowers the solidus temperature thus reducing the temperature existence region of the high temperature phase. The same holds for the Ni-modified
Cu-In eutectoid. This means that tor unidirectional eutectoid decomposition it becomes difficult to avoid melting of the high temperature phase if a steep temperature gradient is to be imposed. In this case melting of the high temperature phase may cause liquid phase segregation and thus the alignment of the eutectoid and/or its composition may be affected in an undesirable way.
lil UNIDIRECTIONAL EUTECTOID DECOMPOSITION AND À.-R RELATIONS
3.1. THE HlSTORY OF UNIDIRECTIONAL DECOMPOSITION
Although eutectoid decomposition is one of the first solid state phase transitions studied - mainly because of rnan's interest in the Austenite Pearlite reaction in the FeC system -little progress was made until the virtues of the isothermal reaction technique became evident from the workof Bain and his collaborators 1930 (53.). Unidirectional eutectoid decomposition, however, is relatively new. The interest in unidirectional eutectoid
decomposition arose when many investigators had already prod!Jced aligned composites by directional control of eutectic solidification. The first to report several unsuccessful attempts to produce al igned composites by unidirectional eutectoid decomposition of Fe-C-austenite was Kraft 1966 (12.). Bolling and Richman (21.) moved a cylinder of austenite through an extremely steep temperature gradient 2500°C/cm thereby achieving a relatively planar transformation front. However, many finecolonies were produced and only a smal I fraction of the pearlite appeared to be aligned along the temperature gradient.
The first to successfully achieve unidirectional eutectoid decomposition producing aligned composites was Carpay 1970 (20.). His experiments concerned the eutectoid systems Co-Si, Cu-Al, Cu-In and Ni-In. The lamellae occurring upon decomposition were parallel to the temperature gradient throughout the samples and parallel toeach other over mms (Co-Si, Ni-In, Cu-Al) or cm (Cu-In). For the eutectoid decomposition of Co3Si, he found an experimental relationship between interlamellar spacing À. (varying between 0.1 and 1 fJm) and the decomposition rate R, which appeared to be: À.4 R =Constant. At relatively high rates (depending on the system) the parallelism of the lamellae was lost. Livingstone 1970 (22.) improved the method of unidirectional eutectoid decomposition of the Cu-Al eutectoid, by exploiting the fact that the high temperature {3-phase of eutectoid composition has a congruent melting point. Thus he employed a technique of unidirectional solidification and unidirectional eutectoid decomposition in the same run.
In this way he obtained single crystalline (3 which - because of the lack of transverse
grain-boundaries- decomposed into eutectoid colonies, that extended along the entire lenght of the sample. lt was suggested by Livingston et al. 1970 (54.) that high speed directional control of eutectoid decomposition might be used to produce aligned composites finer than those attainable by eutectic solidification. This prediction, however, overlooked the fact that the maximum velocities, at which such solid state phase transformation frontscan advance, are much lower than the maximum attainable solidification rates. For the Cu-Al eutectoid decomposition, Asundi and West 1966 (40.) observed a maximum velocity of 1.25 x 10--4 cm/sec. This is in accordance with the findingsof Carpay 1970 (20.) and Livingston 1970 (22.), who both reported that only decomposition rates below 1.4 x 10--4 cm/sec produced aligned composite structures in Cu-11.8 wt
%
Al. However, Carpayremarked in his first paper on unidirectional eutectoid decomposition 1970 (20.) that the maximum pulling rate, still giving rise to aligned lamellae, increased with increasing decomposition temperature (Table 111.1).
