9

Thermal Equilibrium
Diagrams

9.10 The subject of phase equilibrium was briefly discussed in Chapter

7 to introduce the reader to this very important field of study in as
painless a manner as possible. The reader will have acquired a general
impression of the significance of equilibrium diagrams, since no useful
examination of the microstructures of alloys is possible without reference
to the appropriate diagram, which, for this reason, becomes one of the
metallurgists most useful 'tools'. With its aid he can predict with precision
the structure of a given alloy at any given temperature provided that the
alloy's thermal-treatment history is known. In the thermodynamical sense,
'equilibrium' refers to the state of balance which exists, or which tends to
be attained, between the phases in the structure of an alloy after a chemical
or physical change has taken place. In some cases equilibrium may not be
reached for long periods after the change has begun. In fact, rapid cooling
(or 'quenching') to room temperature may suppress the chemical or physi-
cal change involved to such an extent that equilibrium will never be reached
so long as the alloy remains at room temperature. Reheating (or 'temper-
ing') followed by slow cooling to room temperature will allow the physical
or chemical change to proceed in some degree towards completion; the
extent to which equilibrium is thus attained being dependent upon the
temperature reached, the time the alloy remains at that temperature and
the rate of cooling to room temperature or at whatever other temperature
equilibrium is being studied.
9.11 We have already seen that physical and chemical changes which
take place after solidification in carbon steel do so with relative ease
because the small carbon atoms can move quickly through the crystal
lattice of face-centred cubic iron. Thus, during the process of normalising,
a steel will attain structural equilibrium. In some alloy systems, however,
physical changes which take place after solidification are so sluggish that
it is doubtful if equilibrium is ever attained. Often we have no means of
telling whether or not equilibrium has been reached, even after extremely
slow rates of cooling involving periods of days or even weeks. There is
therefore a tendency these days to refer to these charts as 'constitutional
diagrams' instead of 'thermal-equilibrium diagrams'. The latter term, how-
ever, is still in general use, and we shall employ it throughout this book.
9.12 The thermal-equilibrium diagram is in reality a chart which shows
the relationship between the composition, temperature and structure of
any alloy in a series. Let us consider briefly how these diagrams are con-
structed.
A pure metal will complete its solidification without change of tempera-
ture (3.12), whilst an alloy will solidify over a range of temperature which
will depend upon the composition of the alloy (7.41). Consider, for
example, a number of alloys of different composition containing the two
metals A and B which form a series of solid solutions (Fig. 9.1). For
successive compositions containing diminishing amounts of the metal A,
freezing commences at a\, <z2, «3, etc., and ends at bi, fc, b3, etc. Thus, if
we join all points ai, ai, a?>, etc., we shall obtain a line called the liquidus,
indicating the temperature at which any given alloy in the series will com-
mence to solidify. Similarly, if we join the points £>i, bi, ^3, etc., we have
a line, called the solidus, showing the temperature at which any alloy in
the series will become completely solid. Hence the liquidus can be defined
as the line above which all indicated alloy compositions represent com-
pletely homogeneous liquids, whilst the solidus can be defined as the line
below which all represented alloy compositions of A and B are completely
solid. For temperatures and compositions corresponding to the co-
ordinates of points between the two lines both liquid solutions and solid
solutions can co-exist in equilibrium.
9.13 Such a system as this occurs when the two metals are soluble in

liquid only
m.pt.
OfB
liquidus
liquid
solid
TEMPERATURE

solidus
solid only

m.pi.
of A

time-temperature cooling curves

O°/oB IOO^bB
COMPOSITION
IOO°/oA O^oA

Fig. 9.1 The construction of a simple equilibrium diagram of the solid solution type using
cooling curves of a series of alloys.
each other in all proportions in both the liquid and solid states, as in the
case of copper and nickel or gold and silver. When, on the other hand,
the metals are completely insoluble in the solid state a eutectic type of
equilibrium diagram represents the system (Fig. 9.2). Since the eutectic
part of the structure of any alloy in the series solidifies at the same tempera-
ture, and is the portion of the alloy with the lowest melting point, the line
ACEDB will be the solidus whilst AEB is the liquidus.
9.14 Diagrams of this type can be constructed by using the appropriate
points (obtained from time-temperature cooling curves) which indicate
where freezing began and where it was complete. Thus the liquidus and
solidus temperatures for a number of alloys in the series between 100% A
and 100% B can be obtained and then plotted on a temperature-compo-
sition diagram as indicated in Figs. 9.1 and 9.2. Such cooling curves will
determine the positions of the liquidus and solidus; and if the alloy system
is one in which further structural changes occur after the alloy has solidi-
fied, then the metallurgist must resort to other methods of investigation
to determine the phase boundary lines. These methods include the use
of X-rays, electrical-conductivity measurements, dilatometric methods
(which measure changes in volume) and exhaustive microscopical examin-
ation following the quenching of representative specimens from different
temperatures.
9.15 The construction of equilibrium diagrams, then, is the result of
much tedious experimental work, and, since the conditions under which
the work is carried out can be so variable, it is not surprising that the
values assigned to compositions and temperatures at which phase changes
occur are under constant review by research metallurgists. The reader
will, therefore, find that the equilibrium diagrams printed in books on
metallurgy often differ in small detail. In general, however, the variations
are so small as not to affect the treatment of most of the industrially

t i m e - t e m p e r a t u r e cooling curves liquid only

TEMPERATURE

liquid + solid

solid only

0<#>B COMPOSITION 10O4YoB

IOO°/oA O*yoA

Fig. 9.2 The construction of a simple equilibrium diagram of the eutectic type using cooling
curves of a series of alloys.
important alloys. As an example, the accepted value for the carbon content
of a completely eutectoid carbon steel (7.55) has varied between 0.80 and
0.89% during recent years, whilst the eutectoid temperature (the lower
critical temperature of plain carbon steels) has varied between 698 and
732°C.

The Phase Rule

9.20 In the previous chapter a 'phase' was described as either a pure
metal, a solid solution or some form of intermediate compound so long
as it was a single homogeneous substance. Similarly a single element or
compound can exist in different physical forms, each one a separate phase.
Thus the chemical substance 'H 2 O' can exist as ice, water or water vapour,
each of these conditions constituting a 'phase'. The 'phase structure' of
H2O is governed by the prevailing conditions of temperature and pressure
as indicated in the phase diagram in Fig. 9.3. This diagram shows the effects
which variations of temperature and pressure will have on the stability of
a given phase. The 'fusion curve' is an almost vertical line and this indicates
that variation in pressure has very little effect on the melting point of ice.
The same is true of the effects of pressure on the melting points of alloys
for which reason we can neglect pressure when studying the phase equilib-
rium of liquid or solid alloys.
The 'vaporisation curve' however indicates that the boiling point of water
is very dependent on pressure. Water boils at 1000C when the pressure is
equivalent to 760 mm of mercury. If the pressure is reduced to, say, 300 mm
of mercury then water will boil at about 500C which is the reason why tea
cannot be brewed successfully near the summit of Everest.
PRESSURE (mm Hg)

ICE WATER
fusion

WATER
VAPOUR
sublimation
WATER VAPOUR

TEMPERATURE ( 0 C)

Fig. 9.3 The ice-water-water-vapour phase diagram.

 The 'sublimation curve' shows that at low pressures solid ice can vaporise
—or 'sublime'—without passing through the liquid phase. Some sub-
stances, notably carbon dioxide ('dry ice') and iodine, do this at normal
atmospheric pressures when heated to the 'sublimation temperature'.
9.21 In 1876 Gibbs (8.13) introduced his Phase Rule which predicts
the relationship between the number of phases, the number of components
(or pure chemical substances) and the conditions of temperature and pres-
sure which prevail. The general form of the rule can be written:
P+ F=C+ E

where P = the number of different phases present;

C = the number of components (stable chemical substances);
F = the degrees of 'freedom', or 'variance';
E = the environmental factors (temperature and pressure).

9.22 The implications of this rule can be explained more easily by

reference to examples. Consider point A in Fig. 9.3. Here there is only
one phase in the system (water vapour) and one component (the chemical
compound, H2O) whilst there are two environmental factors (temperature
and pressure) to be considered. Hence, since—
P + F= C + E
1 + F= 1 + 2
ie F= 2

This means that two degrees of freedom are involved so that both tempera-
ture and pressure can be altered independently without altering the number
of phases in the system, which is consequently said to be bivariant.
Now consider the point B. This indicates that water is boiling at 1000C
and 760 mm pressure so we have two phases (water and water vapour)
present at the same time, but still one component (H2O). Applying the
Phase Rule:
2 + F= 1 + 2
or F= 1
This means that only one degree of freedom is available and if tempera-
ture is increased pressure must also be increased in sympathy if the water
is to continue to boil steadily. Conversely if, say, pressure were reduced
then the boiling point of water would decrease in sympathy so that con-
ditions were still represented by a point on the 'vaporisation curve'. Failure
to observe these conditions would mean that one phase would disappear
—either the water would cease to boil or it would turn very quickly to
steam. The system represented by B is said to be univariant.
At the triple point T all three phases, ice, water and water vapour can
exist together. Hence:
 3 + F= 1+ 2
or F = O

In this case there is no degree of freedom and if either temperature or

pressure are altered then at least one of the phases will disappear. This
system is invariant.
9.23 We will now consider the Phase Rule as it applies to a study of
metallic alloys. Here the components in the system are pure metals or
elements from which the alloy is made, whilst the phases produced will
be either liquids, solid solutions, intermediate compounds or, rarely, pure
metals.
We are usually dealing with alloy systems in the liquid and solid states
at atmospheric pressure. Since small variations in atmospheric pressure
will have virtually no effect on the stability of phases we can neglect the
effect of pressure and standardise at 'one atmosphere'. As we are now
dealing with only one variable environmental factor, temperature, we can
adapt Gibbs' Phase Rule so:

P + F=CH-I

Moreover since one environmental factor has disappeared we can still

express a two-component system in a two-dimensional diagram, variation
in composition replacing variation in pressure. We will now examine such
a system (Fig. 9.4).
Point A represents a completely homogeneous molten alloy containing
a single liquid phase composed of the two components bismuth and cad-
mium. Hence:

1 + F =2+ 1
or F = I
TEMPERATURE

liquid

Bi + liquid Cd + liquid

Bi 4- Cd

COMPOSITION (OOqbCd
iooyo OfcBi

Fig. 9.4 The bismuth-cadmium phase (or thermal equilibrium) diagram.

There are two degrees of freedom so that both temperature and compo-
sition may be varied independently without altering the single-phase
structure.
On cooling the alloy to B the liquid becomes saturated with cadmium
so that solid cadmium begins to form. There are now two phases present
and:
2 + F =2+ 1
and F = 1

Thus there is one degree of freedom and if the temperature is altered the
composition of the remaining liquid must alter so that B remains on CE.
At E solid cadmium and solid bismuth co-exist with remaining liquid so
that there are three phases and:

3 + F= 2 + 1
and F = 0

Therefore the system at E is invariant and any alteration of temperature

or composition will lead to the disappearance of at least one of the phases.
Conversely if the alloy of composition indicated by E is cooling, then the
temperature must remain constant until the liquid has crystallised com-
pletely. That is, a eutectic crystallises at a constant temperature.
9.24 In the interpretation of thermal equilibrium diagrams it is often
more convenient to make use of a set of definitions and rules which are
derived from Gibbs' Phase Rule and also from some observations made
by the French metallurgist A. Portevin:

(a) The areas of the diagram are called 'phase fields', and on crossing
any sloping boundary line from one field to the next, the number of
phases will always change by one, ie two single-phase fields will always
be separated by a double-phase field containing both of the phases. In
a binary system three phases can only exist together at a point, such as
the eutectic point E in Fig. 9.4.
(b) In Fig. 9.5 the point P represents some particular alloy mixture in
both composition and temperature. The diagram indicates that at P both
liquid and solid will exist together. If a 'temperature horizontal' (usually
known as a tie-line) is drawn through P the composition of the solid is
given by X, and the composition of the liquid solution in equilibrium
with it by Y. P itself represents the overall composition of the mixture.
(c) The relative amounts of both liquid and solid at P are given by the
relative lengths PX and PY-

Weight of solid solution Weight of liquid solution

(composition X) x PX = (composition Y) x PY
 liquid

liquid 4- solid
TEMPERATURE

solid

COMPOSITION
Fig. 9.5.

This is generally referred to as the Lever Rule since it is as though we

were taking moments about P as the fulcrum. The relationship can be
expressed in more convenient form:
Weight of solid solution (composition X) PY
Weight of liquid solution (composition Y) PX
(d) A phase which does not occupy a field by itself, but appears only
in a two-phase field, is either a pure metal or an intermediate phase
TEMPERATURE

COMPOSITION

Fig. 9.6.
 (usually an intermetallic compound) of invariable composition. In fact
the phase 'field' will be infinitely narrow such as is represented by the
100% bismuth ordinate in Fig. 9.4 or the Mg2Sn ordinate in Fig. 9.15.
(e) If we draw a vertical line on a diagram this will represent the
composition of a given alloy (X in Fig. 9.6). If, on following along this
line, a phase boundary is crossed (at O) the number of phases will change
by one at this point, ie a phase (|3) will be either precipitated or absorbed.
(In this instance (3 will be absorbed when the alloy X is heated above O
and precipitated again when the alloy X cools below O.)

We will now consider the main types of thermal equilibrium diagram

which are of use in studying metallic alloy systems. Generally a useful alloy
will only be formed when the two metals are completely soluble in each
other in the liquid state; but in some instances the two metals are only
partially soluble as liquids. We will begin with a brief study of one such
case.

Case I—Two metals which are only partially

soluble in each other in the liquid state
9.30 Fig. 9.7 represents the phase equilibrium conditions for the zinc-
lead system above 5000C. As in all of these equilibrium diagrams the base
line shows the composition varying between 100% of one component at
one side of the diagram to 100% of the other component at the opposite
side. In most diagrams the percentages are by weight but in some cases it
is more relevant to indicate percentages in terms of numbers of atoms.
The ordinate represents temperature. Thus point P represents a mixture
consisting of 70% by weight of zinc and 30% by weight of lead at a tempera-
ture of 7000C.
9.31 The diagram (Fig. 9.7) shows that at 5000C molten zinc will dis-
solve no more than 2% lead (point A) giving a single solution of lead in
zinc. At the same temperature molten lead will dissolve about 5% zinc
(point B) thus giving a single solution of zinc in lead. However, any alloy
at 5000C which has a composition given by some point between A and B
will consist of two separate layers of two different liquid solutions—the
upper layer will be zinc containing 2% dissolved lead (composition A) and
the lower layer will be lead containing 5% dissolved zinc (composition B).
9.32 The slope of the line ADJC indicates that as the temperature rises
lead becomes increasingly soluble in molten zinc. Similarly the slope of
BEC shows that, under similar conditions, the solubility of zinc in molten
lead also increases. These two solubility curves meet at 798°C indicating
that at temperatures above this, the two metals become completely soluble
in each other so that a single liquid phase will be present in the mixture
whatever its overall composition.
9.33 Consider a molten mixture consisting of 70% zinc and 30% lead
(composition H). At 5000C this will consist of a saturated layer of lead in
 °C ONE LIQUID

TWO LIQUIDS

KX)<*>Pb
O°A>Zn
COMPOSITION {%> by wt.)

Fig. 9.7 The zinc-lead thermal equilibrium diagram above 5000C.

zinc (composition A) floating on top of a saturated layer of zinc in lead

(composition B). That is, it is a two-phase mixture.
If the temperature is now raised the amount of lead which dissolves in
zinc increases as does the amount of zinc which dissolves in lead. Thus at
7000C zinc will now dissolve 20% lead (composition D) whilst lead will
dissolve 20% zinc (composition E). Using the Lever Rule we can calculate
the relative proportions by weight of each layer:

Weight of zinc-rich layer Weight of lead-rich layer

(composition D)xDP= (composition E) x PE
Weight of zinc-rich layer (composition D) PE
Weight of lead-rich layer (composition E) DP
Substituting values from the diagram:
Weight of zinc-rich layer (80%Zn-20%Pb) _ 80-30
Weight of lead-rich layer (80%Pb-20%Zn) ~ 30-20
_5
~T
As the temperature of this 70-30 mixture is raised still further to 752°C
the two liquid phases merge into one, as the two metals become completely
soluble in each other at this composition.
Case II—Two metals mutually soluble in all
proportions in the liquid state becoming
completely insoluble in the solid state
9.40 It is extremely doubtful whether such a situation as this really exists
in practice since most solid metals appear to dissolve quantities, however
small, of other metals. However, in the case of bismuth and cadmium the
mutual solid solubility is so small that we can neglect it for the sake of our
argument and assume that the metals are completely insoluble in the solid
state. Bismuth, a heavy, brittle, slightly pink metal, is one of the least
'metallic' of metals as indicated by its position near to the non-metals in
the Periodic Table. It forms a complex rhombic-type crystal structure in
which the metallic bond is partly replaced by a mixture of covalent bonds
and van der Waals forces. Cadmium on the other hand crystallises as a
normal CPH structure so that it is not surprising that these two metals with
such unlike lattice structures do not form mixed crystals but separate out
almost completely on solidification. The bismuth-cadmium thermal-
equilibrium diagram is shown in Fig. 9.8.
9.41 Let us consider the solidification of an alloy of composition x, ie
containing about 80% cadmium and 20% bismuth. When the temperature

THE BISMUTH-CADMIUM
EQUILIBRIUM DIAGRAM.

LIQUID

LIQUID + LIQUID + CADMIUM

BISMUTH

BISMUTH + CADMIUM

(Bi + EUTECTIC) (Cd + EUTECTIC)

COMPOSITION PER CENT

BISMUTH
THE RELATIVE RICHNESS OF THE
CADMIUM
LIQUIO PHASE IN CADMIUM IS INDICATED
EUTECTIC BY THE DISTANCE APART OF THE
SHADING LINES.
LIQUID

Fig. 9.8 The bismuth-cadmium equilibrium diagram.

falls to T, crystal nuclei of pure cadmium begin to form in accordance with
rule (b) stated in 9.24. (The temperature horizontal or tie-line, T, cuts the
liquidus at the chosen composition, x, and the other phase boundary is the
100% cadmium ordinate.) Since pure cadmium is deposited, it follows
that the liquid which remains becomes correspondingly richer in bismuth.
Therefore the composition of the liquid moves to the left—say, to x\—
and, as indicated by the diagram, no further deposition of cadmium takes
place until the temperature has fallen to Ti. When this happens more
cadmium is deposited, and dendrites begin to develop from the nuclei
which have already formed. The growth of the cadmium dendrites, on the
one hand, and the consequent enrichment of the remaining liquid in bis-
muth, on the other, continues until the temperature has fallen to 1400C.
The remaining liquid then contains 40% cadmium and 60% bismuth, ie
the eutectic point E has been reached.
At this point the two metals are in equilibrium in the liquid, but, due to
the momentum of crystallisation, the composition swings a little too far
past the point E, resulting in the deposition of a little too much cadmium.
In order that equilibrium shall be maintained, a swing back in composition
across the eutectic point takes place by the deposition of a layer of bismuth.
In this way the composition of the liquid oscillates about E by depositing
alternate layers of cadmium and bismuth, whilst the temperature remains
at 1400C until the remaining liquid has solidified. Thus the final structure
will consist of primary crystals of cadmium which formed between the
temperature T and 1400C, and a eutectic consisting of alternate layers of
cadmium and bismuth which formed at 1400C.
9.42 Had the original liquid contained less than 40% cadmium, then
crystals of pure bismuth would have formed first, causing the composition
of the remaining liquid to move to the right until ultimately the point E
was reached as before, and the final liquid contained 40% cadmium and
60% bismuth. This remaining liquid would solidify as eutectic in the
manner already described.
9.43 If the original liquid contained exactly 40% cadmium and 60%
bismuth at the outset, then no solidification whatever would occur until
the temperature had fallen to 1400C. Then a structure composed entirely
of eutectic would be formed as outlined above.
9.44 In all three cases mentioned, the eutectic part of the structure will
be of constant composition and will always contain 40% cadmium and 60%
bismuth. Any variation either side of this in the overall composition of the
alloy will be compensated for by first depositing appropriate amounts of
either primary cadmium or primary bismuth, whichever is in excess of the
eutectic composition. It is important to realise that there is no question
of solid solubility existing in any way in the final structure, whatever its
composition. With the aid of a microscope, we can see the two pure metals
cadmium and bismuth as separate constituents in the microstructure. In
other words, this is a case of complete insolubility in the solid state.
 0
C

liquid
NiAI3 + liquid-
Al + liquid
NiAI3

Al + NiAI3

(Al + eutectic) (NiAI3 • eutectic)

COMPOSITION (°/owt.)
Plate 9.1 Part of the aluminium-nickel thermal equilibrium diagram which includes the
eutectic system between aluminium and the intermetallic compound NiAI3.
The photomicrographs of two representative alloys in the cast condition are shown. The 9 7 -
3 alloy contains approximately equal amounts of primary aluminium crystals (the light areas)
and eutectic. The structure of the 90-10 alloy consists of primary NiAI3 crystals (dark) and
eutectic. Note the regular 'geometric' shape of the primary NiAI3 crystals—this is often a
feature of intermetallic compounds.
The light aluminium layers in the lamellar eutectic are much broader than those of the dark
NiAI3 because the eutectic composition contains only 5.7% nickel.

Case Ill—Two metals, mutually soluble in all

proportions in the liquid state, remain mutually
soluble in all proportions in the solid state
9.50 A number of pairs of metals fulfil these conditions and are generally
those in which:
(i) the two atoms concerned are compatible in terms of the atomic
size factor (8.23);
 (ii) the electrochemical properties of the two metals are very similar;
(iii) the crystal structures of the two pure metals are similar in pattern.
If these conditions are fulfilled it is reasonable to expect atoms of the
second metal to be able to replace atoms of the first in all proportions to
form a disordered solid solution. Such is the case in the alloy systems: Ag-
Au; Ag-Pd; Au-Pt; Bi-Sb; Co-Ni; Cu-Ni; Cu-Pt; Fe-Pt; Ni-Pt;
and Ta-Ti. (In some of these systems further phase changes occur in the
solid state due to polymorphic transformations (3.14)). Since the copper-
nickel alloys are the only ones of the group mentioned above which are of
use in engineering it is proposed to deal with this system. The cop-
per-nickel thermal-equilibrium diagram is shown in Fig. 9.9.
Again this is a simple type of equilibrium diagram, and since, as in the
cadmium-bismuth system, no transformations take place in the solid, the
diagram consists of two lines only—the liquidus and solidus. Above the
liquidus we have a uniform liquid solution for any alloy in the series, whilst
below the solidus we have a single solid solution for any alloy, though in
the cast condition, as we shall see, the solid solution may vary in concen-
tration due to coring. Between the liquidus and solidus both liquid and
solid solutions co-exist.
9.51 Applying the Phase Rule to that portion of the diagram above the
liquidus:

THE COPPER-NICKEL EQUILIBRIUM DIAGRAM.

0
C

LIQUID

LIQUID
SOLID

SOLID

IOO Ni
OC.
COMPOSITION PER CENT

LIQUID NICKEL RICH COPPER RICH

SOLID SOLUTION SOLID SOLUTION

Fig. 9.9 The copper-nickel equilibrium diagram.

The microstructures indicated are those obtained under non-equilibrium conditions of solidi-
fication.
 p + F= C+ 1
1+F =2+1
and F = 2

since there are two degrees of freedom both temperature and composition
can be altered over fairly wide limits without upsetting the stability of the
single phase. However, for a point between the liquidus and solidus, ie in
'liquid + solid' field:

2 + F=2+ 1
and F = 1
There is thus one degree of freedom, and if either temperature or compo-
sition are altered independently of the other, then the system becomes
unstable so that some solid will melt or some liquid will solidify. Thus any
alteration in temperature must be accompanied by a change in overall
composition if the proportions of liquid and solid are to remain the same.
9.52 Consider the freezing of an alloy of composition A (Fig. 9.9), ie
containing 60% nickel and 40% copper. Assume that cooling is taking
place fairly rapidly, such as would be the case during the industrial casting
of slab ingots into cast-iron moulds. Above T0C the alloy will exist as a
uniform liquid solution, but as the temperature falls to T0C, dendrites of
solid solution will begin to form. They will not however be dendrites of
composition A but dendrites of composition B. This composition B is
indicated by drawing a tie-line from the liquidus at A (which represents
the composition and temperature of the original liquid) to the solidus,

Plate 9.2 70-30 cupro-nickel in the chill-cast condition, x 80.

This shows a mass of dendrites of heavily cored solid solution. The lighter cores are nickel-
rich whilst the copper-rich infilling has etched a darker colour.
which it cuts at B. Thus the dendrites which form will contain approxi-
mately 75% nickel (composition B) and, since the original liquid contained
only 60% nickel (composition A), it follows that the remaining liquid will
contain an even lower proportion of nickel. Hence its composition will
move to the left, say to A\.
Solidification will continue when the temperature falls further to T1, and
this time a layer of solid of composition B1 will be deposited. This is less
rich in nickel than the original seed crystals and, as crystallisation proceeds,
successive layers will contain less and less nickel, and consequently more
and more copper, until ultimately the liquid is used up.
Clearly, then, a non-uniform solid solution is formed, and whilst its
overall composition will be 60% nickel and 40% copper, due to the coring
effect the initial skeleton of a crystal will contain about 75% nickel and its
outside fringes only about 50% nickel.
9.53 The situation is further influenced, however, by diffusion (8.23),
which is taking place simultaneously with crystallisation. Due to the fact
that successive layers of the alloy which deposit are richer in nickel than
the remaining liquid, a concentration gradient is set up which tends to
make copper atoms diffuse inwards towards the centre of the dendrite,
whilst nickel atoms move outwards towards the copper-rich liquid. In the
above case we have assumed very rapid cooling so that little or no diffusion
has been possible, and that under these conditions the structure has been
unable to reach equilibrium so that a heavily cored structure forms. Since
we are dealing with non-equilibrium conditions, only a limited amount of
information can be obtained from the equilibrium diagram—in this case a
guide as to why the cored structure develops. If cooling were slow, dif-
fusion could take place extensively so that coring would be slight in the
final crystals. Annealing at a sufficiently high temperature will also permit
diffusion to proceed (8.24), resulting in the formation of almost completely
uniform solid-solution crystals.
9.54 We will now consider the cooling of the same alloy under con-
ditions which are ideally slow so that complete equilibrium is attained at
each stage of solidification. The liquid (composition A, Fig. 9.10) will begin
to solidify at temperature T by depositing nuclei of composition B. This
causes the composition of the remaining liquid to move to the left, but,
due to the prevailing slow cooling, diffusion is able to keep pace with

LIQUID
TEMPERATURE

LIQUID + SOLID

SOLID

COMPOSITION PER CENT

Fig. 9.10.
solidification so that, as the composition of the liquid follows the liquidus
from A to A2, the composition of the solid follows the solidus from B to
B2.
Thus, at some temperature Ti, the composition of the uniform solid
solution is given by B\, whilst that of the remaining homogeneous liquid
in equilibrium with it is given by A\. Since the overall composition of the
alloy is indicated by X (which is the same as A) then, by the Lever Rule:
Weight of solid solution (composition B1) XA1
Weight of remaining liquid solution (composition A1) XB1
At the temperature T2 the last trace of liquid (composition A2) has just
solidified at the crystal boundaries and, by the process of diffusion, its
composition has changed to that of the remainder of the structure which
has adjusted itself to uniform B2 throughout. A and B2 of course represent
the same composition, which is obvious, since a single uniform liquid sol-
ution has been replaced by a single uniform solid solution.
9.55 When two metals form a substitutional solid solution some distor-
tion of the lattice is inevitable due to influences such as different atom size.
In some cases the two metals are only just compatible and although they
are still soluble in all proportions in the solid state a phase diagram of the
type shown in Fig. 9.11 results. A minimum melting point (32.80C) is
produced where liquidus and solidus are congruent. The significance of this
minimum point is that the metals are so incompatible in solid solution that
there is a strong tendency for them to return to the disordered liquid state

0
C

liquid liquid solution

solid

liquid +
solid

solid solution

COMPOSITION (^>wt)
Fig. 9.11 The potassium-rubidium thermal equilibrium diagram.
at as low a temperature as possible. In fact a little more incompatibility
would lead to a complete 'divorce' to the extent that the metals would
separate forming a eutectic, probably of two solid solutions. We shall in
fact deal with this type of system in the next case.

Case IV—Two metals mutually soluble in all

proportions in the liquid state but only partially
soluble in the solid state
9.60 This case is in effect intermediate between the two dealt with in
Cases II and III and the thermal equilibrium diagram is, in consequence,
a sort of hybrid of these two. It represents a stage where incompatibility
has exceeded that suggested in Fig. 9.11. As in the cadmium-bismuth
system (where the metals are completely incompatible in the solid state)
a eutectic is formed, but in this case it is a eutectic of two solid solutions
instead of two pure metals. It is in fact the 'general' case and purists may
argue that, as such, it should have been dealt with before either of the
preceding particular cases. The systems have, however, been dealt with in
order of complexity rather than logical classification in order to assist the
student who may be encountering the subject for the first time.
The only new feature of this system, as compared with those foregoing,
is that we have phase boundaries occurring below the solidus, indicating
that phase changes can take place in the solid. On the tin-lead thermal-
equilibrium diagram (Fig. 9.12), which we shall use as an example of this
type of system, such phase boundaries as these are indicated by the lines
BC and FG. A line such as BC or FG is termed a solvus and indicates a
change in solubility of one metal in another in the solid state. Part of the
tin-lead system was used in Chapter 7 as an introduction to equilibrium
diagrams. We will now consider it in its entirety.
9.61 As the diagram shows, we have two solid solutions in the system,
since:

(a) Tin will dissolve up to a maximum of 2.6% lead at the eutectic

temperature, forming the solid solution a.
(b) Lead will dissolve up to a maximum of 19.5% tin at the eutectic
temperature, giving the solid solution (3.
(The different phases present in an alloy series are generally lettered from
left to right across the diagram, using letters of the Greek alphabet for
convenient reference.)
The slope of the phase boundaries BC and FG indicates that both the
solubility of lead in tin (a) and the solubility of tin in lead (|3) decrease as
the temperature falls. This is a normal phenomenon with solid solutions
as it is with liquid solutions. In Fig. 9.13 the solubility of a salt in water at
different temperatures is compared with the solid solubility of one metal
in another at different temperatures. In each case solubility increases as
 LIQUID
LIQUID

LIQUID +• /S>

LIQUID + <y.

TIN-RICH SOLID

a + /3 LEAD-RICH SOLID

COMPOSITION PER CENT

EUTECTIC

CORED O. CORED(^ + ALL EUTECTIC CORED /3 + CORED /3

EUTECTIC OF U)3 'EUTECTIC

Fig. 9.12 The tin-lead equilibrium diagram

The microstructures indicated are those obtained under non-equilibrium conditions of solidification.
 solubility solvus
curve

TEMPERATURE
TEMPERATURE

unsaturatcd unsaturatcd
liquid solid
solution solution j
(cC) /

saturated liquid solution saturated solid solutionWC)

+ solid s a l t + solid B

undissolved
salt metal B

10O4Yo water °A> salt IOOfcmctalA % metal B

Fig. 9.13 An analogy between the liquid solubility of salt in water and the partial solid
solubility of one metal in another.

temperature rises and saturated solutions are formed in a similar manner.

X and X\ each represent a mixture which at that temperature is an unsatu-
rated solution, ie it could dissolve more solute. As the temperature falls to
S (Si) the solution (liquid or solid) reaches a saturation point so that solid
salt (or metal B) begin to precipitate. At T (Ti) a considerable quantity Q
(Qi) of solute (either salt or metal B) is still in solution but as the tempera-
ture falls to R (Ri) this is reduced to a very small amount in each case—
Y (Yi) as salt (or metal B) is progressively precipitated from solution.
9.62 In the case of the tin-lead alloys the sloping boundaries BC and
FG indicate that changes take place in the solid. Thus as the temperature
falls:
(i) lead becomes less soluble in solid tin (a) along BC;
(H) tin becomes less soluble in solid lead (|3) along FG.
Several different structures may be formed, depending upon the alloy
composition. For example, let us consider:
(a) An alloy of composition X (70% lead-30% tin). This will begin to
solidify when the temperature falls to T and dendrites of composition Y
will deposit. The alloy continues to solidify in the manner of solid sol-
ution until at 183°C the last layer of solid to form will be composition F
(80.5% lead-19.5% tin) and the remaining liquid will be of composition
E (the eutectic composition with 38% lead and 62% tin).
The remaining liquid now solidifies by depositing, in the form of a
eutectic, alternate layers of a and |3, of the compositions B and F respect-
ively. If this structure now cools slowly to room temperature the compo-
sitions of the solid solutions a and |3 will follow the lines BC and FG, ie
the solid solution a will become progressively poorer in lead and the
solid solution (3 will become poorer in tin, until at, say, 1000C a will
contain less than 1% lead and (3 will contain less than 10% tin. The
 (i)

(ii)

(iii)

Plate 9.3 Representative tin-lead alloys in the cast condition, x 100.

(i) 75 tin-25 lead. Primary crystals of a (light) in a matrix of eutectic consisting of layers of
a (light) and p (dark).
(ii) 62 tin—38 lead. Completely eutectic in structure—layers of a (light) and P (dark).
(iii) 50 tin-50 lead. Primary crystals of p (dark) in a matrix a + p eutectic as in (i) and (ii).
Note the small 'islands' of a within the primary |3 crystals in (iii). These were precipitated as
the alloy cooled slowly from 183°C to ambient temperature and the P changed in composition
along the steeply sloping solvus FG (Fig. 9.12).
 proportions of a and (3 will also vary from EFIEB at 183°C to E'JIE'H
at 1000C. In the same way, provided we assume that the structure cooled
slowly enough to allow the primary dendrites of (3 to reach a uniform
composition F, these dendrites will now alter in composition to / at
1000C, and in so doing precipitate some a of composition H in order to
adjust their own composition. The a precipitated will join that already
present in the eutectic.
(b) If the alloy contained, say, 95% lead, and was cooled slowly
enough to prevent coring, solidification would be complete at P and a
uniform solid solution |3 would result. On continuing to cool slowly,
further solid changes would begin at Q with the precipitation of small
amounts of a at the grain boundaries of the (3. This a would increase in
amount as the temperature fell, and the |3 became progressively poorer
in tin. Hence the final structure would consist of crystals of uniform (3
containing about 98% lead with small amounts of a precipitated at the
crystal boundaries.
(c) If the original alloy contained more than 98% lead and was cooled
slowly, the structure would remain entirely (3 throughout, after its solidi-
fication had been completed at R.
9.63 In actual practice it is unlikely that cooling would be slow enough
to prevent some coring from taking place. Since the initial |3 dendrites
would then be relatively rich in lead, this could lead to the formation of
small amounts of a at the (3 grain boundaries in (c) and more than the
expected amount of a in (b). In case (c) this a would be absorbed on
annealing. 'Solution annealing' is an important process used in the treat-
ment of many alloys to absorb some constituent which has been precipi-
tated due to coring. After treatment the alloy will be soft and ductile and
able to receive cold-work. Solution annealing is also an integral part of
most strengthening processes based on precipitation hardening, the prin-
ciples of which will be discussed later (9.92).

Case V—A system in which a peritectic

transformation is involved
9.70 Sometimes in an alloy system two phases which are already present
interact at a fixed temperature to produce an entirely new phase. This is
known as a peritectic transformation. The term is derived from the Greek
'peri', which means 'around', since, during the transformation process, the
original solid phase becomes surrounded or coated by the transformation
product. Frequently one of the interacting phases is a liquid though this is
not a precondition and sometimes two solids will take part—albeit very
slowly—in a peritecto/d transformation. A peritectic transformation occurs
between the phase 6 and the remaining liquid (producing austenite) in
the iron-carbon system (11.20) but here we will consider the peritectic
transformation which takes place in the platinum-silver system, part of
which is shown in Fig. 9.14.
 0
C liquid
CC

liquid

CC + liquid

OC

S +
liquid
OC + S

COMPOSITION (°/o)

Fig. 9.14 The platinum-silver thermal equilibrium diagram.

(i) A slowly cooled alloy containing about 75% silver will first begin to solidify as uniform a
(ii) Below 1185°C these a crystals will become coated with 6 due to peritectic interaction
with the remaining liquid, (iii) The original a will be transformed slowly to 6 as the temperature
falls, (iv) If cooling is extremely slow all a will be absorbed and the remaining liquid will
solidify as uniform 6. (v) In practice cooling will generally be so rapid as to 'trap' some a
'cores' in the final structure, producing a non-equilibrium structure.

9.71 If we ignore coring effects, that is, assume a very slow rate of
cooling, the peritectic transformation will occur only in those platinum-sil-
ver mixtures containing between 12 and 69% silver. Let us first consider a
liquid of composition x, ie containing about 25% silver. This will begin to
solidify by depositing dendrites of the solid solution a of composition a. If
the rate of cooling is slow enough for diffusion to remove the effects of
coring, by the time the temperature has fallen to 1185°C the structure will
consist of a dendrites containing 12% silver (composition P) and a remain-
ing liquid containing 69% silver (composition R), since the composition of
a will change as it follows the solidus SP. At this stage:
Weight of a (composition P) X1R
Weight of remaining liquid (composition R) X1P

9.72 At 1185°C, the peritectic temperature, the a dendrites begin to

interact with the remaining liquid and form a new solid solution, 6, ie

a + liquid ^ 5

This solid solution b contains 55% platinum-45% silver (composition Q).

However the overall composition of the alloy is x\ (75% platinum-25%
silver) and it follows that not all of the a will be used up, since it is in
excess of that required to produce a structure entirely of 5. Therefore, all
of the liquid is used up and some a remains, so that the final structure will
consist of a containing 12% silver (composition P), and 5 containing 45%
silver (composition Q). These phases will be in a ratio given by:
Weight of a (composition P) X1Q
Weight of 3 (composition Q) X1P
9.73 Let us now consider a liquid alloy of initial composition y. This
will begin to solidify by depositing crystals of a of composition ai, and if
we again neglect coring, the crystals will gradually change in composition
to P when the temperature has fallen to 11850C, and again the composition
of the remaining liquid will be R. At this stage—
Weight of a (composition P) ^1 R
Weight of remaining liquid (composition R) yxP

When the peritectic transformation takes place, we shall find that this time
there is an excess of liquid because yi lies between Q and R. The a will
therefore be completely used up, and just below 1185°C we shall be left
with the new solid solution, 8, and some remaining liquid in the ratio—
Weight of S (composition Q) yx R
Weight of remaining liquid (composition R) yx Q
As the temperature continues to fall slowly, this remaining liquid will
solidify as 5, which will change in composition along Qy2. At y2 solidifi-
cation will be complete and the structure will consist of crystals of uniform
5 of the same composition as that of the original liquid. In practice it is
unlikely that the rate of cooling would be slow enough to permit equilib-
rium being reached. Consequently, though the equilibrium structure is
indicated as being completely b (in the case of alloy y), remnants of the
original a dendrites may be present in the rapidly-cooled structure. The
original a dendrites, produced at temperatures above 1185°C, became
coated with a protective envelope of 8 as the interaction with the liquid
began and subsequent diffusion was far too slow to promote dispersal of
the remaining a.

Case Vl—Systems containing one or more

intermediate phase
9.80 Frequently two metals will not only form limited solid solutions with
each other but, in other suitable proportions, will also interact to form one
or more intermediate phases which may vary in nature from secondary
solid solutions to true intermetallic compounds. The formation of any
intermediate phase further complicates the thermal equilibrium diagram
by introducing extra boundary lines and since some alloy systems include
a number of intermediate phases, formed at different compositions of the
two metals, the resulting diagram becomes increasingly complex.
9.81 We will first examine the tin-magnesium system (Fig. 9.15) as an
example containing a single intermediate phase—in this case the intermet-
allic compound Mg2Sn (containing 29.1% magnesium). This system is rep-
resented by two simple equilibrium diagrams joined at B. The section AB
represents a eutectiferous series of the substances pure tin and Mg2Sn.
Here no solid solubility is involved and section AB can therefore be inter-
preted in the manner of Case II (the bismuth-cadmium system) dealt with
in 9.40. The section BC is a eutectiferous series of the Case IV type (9.60)
involving Mg2Sn and the solid solution a. If we consider the complete
system extending from pure tin to pure magnesium then, clearly, these two
sections become joined as shown, forming a single diagram. Since Mg2Sn
is an intermetallic compound of single fixed composition its 'phase field' is
represented on the diagram by a vertical line at 29.1% magnesium. A
particularly interesting feature of this diagram is that pure Mg2Sn has a
melting point (778°C) which is higher than those of either of the constituent
pure metals. This is an indication of the stability and strong chemical
bonding in this valence-type compound.
9.82 We will now consider the solidification of a representative alloy
in the series, ie alloy 'x' (60% tin-40% magnesium). This will begin to
solidify when the temperature falls to T and the tie-line from x cuts the
next phase boundary at y which represents the pure intermetallic com-

0
C liquid
liquid
Mg2Sn
OC

OZ + liquid
liquid

liquid
liquid

COMPOSITION (%>)

Fig. 9.15 The tin-magnesium thermal equilibrium diagram, consisting essentially of two
eutectiferous series—tin-Mg 2 Sn and Mg2Sn-magnesium.
pound Mg2Sn. Therefore crystals of Mg2Sn begin to form. Since these
crystals are relatively tin-rich (70.9% tin) the composition of the remaining
liquid moves to the right to 'x{ and as the temperature falls to T\ more
Mg2Sn solidifies. Mg2Sn continues to solidify in this way until at 5600C the
remaining liquid contains 63.6% magnesium-36.4% tin—the composition,
E, of the Mg2Sn +a eutectic.
The remaining liquid then solidifies as eutectic at 5600C by depositing
alternate layers of Mg2Sn and the solid solution a, containing 85% mag-
nesium-15% tin (composition S). When completely solid the structure will
consist of primary crystals of the intermetallic compound Mg2Sn sur-
rounded by a eutectic of Mg2Sn and a. If the solid alloy is then allowed to
cool slowly below 5600C, a will change in composition along the solvus line
SL by depositing particles of Mg2Sn which will join the Mg2Sn constituent
already present in the eutectic.
9.83 The large size of lead atoms compared with those of many other
metals contributes to the fact that lead forms relatively few solid solutions.
Thus lead and gold are to all intents and purposes insoluble in each other
in the solid state but, due to differences in the electrochemical properties
of the metals two intermediate phases, Au2Pb and AuPb2 are formed at
the appropriate compositions. These two intermediate phases are produced
as a result of peritectic interactions and the resultant phase diagram is
shown in Fig. 9.16.
9.84 Consider an alloy containing 40% lead. This will begin to solidify

°c

liquid

Au + liquid

Au1Pb + liquid

COMPOSITION (o/o)

Fig. 9.16 The gold-lead thermal equilibrium diagram. Two intermediate phases are pro-
duced as a result of peritectic transformations.
at about 5000C (T) by forming dendrites of gold, but on cooling to 418°C,
the first peritectic temperature, these gold dendrites undergo a peritectic
interaction with the remaining liquid to produce crystals of the intermedi-
ate phase, Au2Pb:

Au + liquid ^ AU2 Pb

These crystals of Au2Pb continue to grow as the temperature falls, but at

254°C a further peritectic interaction occurs resulting in the formation of
another intermediate phase AuPb2:
Au2Pb + liquid ^± AuPb2
During this second transformation all of the liquid is used up first and some
Au2Pb remains along with the new phase AuPb2. Assuming that the system
has cooled slowly to room temperature these two phases will be present
in the ratio:
Weight OfAu2Pb _ BX
Weight of AuPb2 ~~ ~AX
_ 67.8-40
" 40 - 34.4
_27.8

_5
~T
9.85 An alloy containing 70% lead will begin to solidify at about 2800C
(Ti) by depositing crystals of Au2Pb but at 254°C a peritectic interaction
results in the formation of AuPb2. Here all of the Au2Pb is used up and
the remaining liquid continues to solidify as the new phase AuPb2 until at
215°C the remaining liquid is of composition E (85% lead-15% gold). A
eutectic consisting of alternate layers of AuPb2 and pure lead is then
formed so that the final structure consists of primary AuPb2 in a eutectic
of AuPb2 and lead.
In both of these cases it has been assumed that cooling has been slow
enough to permit equilibrium to be attained throughout so that peritectic
transformations proceed to completion. In practice this may not be so.
Thus in the 40% lead alloy more than the indicated quantity of Au2Pb may
be present in the microstructure whilst in the 70% lead alloy some Au2Pb
cores may be present even though the diagram indicates a AuPb2-lead
structure.
9.86 Sometimes an intermediate phase is of variable composition. Thus
the phase (3 in the tantalum-nickel phase diagram (Fig. 9.17) is based on
TaNi3 but it may be thought of as being able to dissolve either excess
tantalum or excess nickel so that the phase field occupies width on the
diagram.
 0 /3 • liquid
C

OC + liquid

COMPOSITION (%>)
Fig. 9.17 Part of tantalum-nickel thermal equilibrium diagram. |3 is an intermediate phase
which varies in composition about TaNi 3 .

Precipitation from a Solid Solution

9.90 When the temperature of a solid solution falls such that it reaches
a state of saturation, any further fall in temperature will lead to the precipi-
tation of some second phase. This phenomenon was mentioned in 9.61 and
is similar in principle to the precipitation which takes place from saturated
liquid solutions when these continue to cool.
Precipitation from a solid solution, however, takes place much more
sluggishly than from a liquid solution because of the greater difficulty of
movement of the solute atoms in a solid solution, particularly if it is of the
substitutional type. We must, therefore, consider carefully the relationship
between the rate of cooling of a saturated solid solution and the extent
to which precipitation can take place—and consequently equilibrium be
attained—either during the cooling process or subsequently.
9.91 Precipitation under Equilibrium Conditions Fig. 9.18 rep-
resents part of a system such as was described in 9.60. Here metal B is
partially soluble in metal A in the solid state forming the solid solution a.
Any B in excess of solubility at any given temperature is precipitated,
under equilibrium conditions, as the phase (3, which in this case may be
either a solid solution or an intermediate phase*.
Consider the slow cooling of some alloy of composition X from the
temperature Tn. At Tu the solid solution a is unsaturated and this state of
affairs prevails until the temperature falls to Ts. Here the solid solution
* Since the remainder of the diagram is not given we cannot know whether (3 is a solid solution as
indicated in 9.60 or some intermediate phase as indicated in the various examples discussed in 9.80.
 liquid

unsaturated CC
CC + liquid solid solution
containing X°/o
of 'B' dissolved."

UJ
CC
3

< OC OC is now saturated

CC
UJ
and A has begun to
a precipitate.
UJ
K-
CC + /3

A continues to
precipitate and
oC now contains
only X1^o of'B'in
solid solution.
METAL 'S" °/o

Fig. 9.18 Changes in partial solid solubility with temperature.

reaches saturation and when the temperature falls below T8, precipitation
may begin. Random nucleation takes place at the grain boundaries as well
as on certain crystallographic planes in the a matrix, and nuclei of [3 begin
to form. As the alloy cools, precipitation of |3 continues, and since |3 is rich
in B, the composition of a changes progressively along the solvus line PQ.
Because (3 is richer in B than is a, B must diffuse through a in order to
reach the growing nuclei of (3. This tends to reduce the concentration of B
rather more rapidly in that a near to the (3 nuclei than in those regions of
a which are far away from any nuclei. The concentration gradient is,
however, too small to cause rapid diffusion of the B atoms, and, since the
temperature is falling, regions of a far from any (3 nuclei ultimately become
supersaturated with excess B. Consequently more (3 nuclei will form and
begin to grow.
Precipitates which form under equilibrium conditions in this manner are
generally non-coherent. That is, the new phase which has formed has a
crystal structure which is entirely its own and is completely separate from
the surrounding matrix from which it was precipitated. Its strengthening
effect on the alloy as a whole is somewhat limited (8.62) and generally
much less than that resulting from the presence of a coherent precipitate,
the formation of which will be dealt with in the next section.
Assuming that the temperature has fallen to TR and that the precipitation
of (3 has taken place under equilibrium conditions, the structure will consist
of non-coherent particles of (3 (usually of such a size as can be seen easily
using an ordinary optical microscope) in a matrix of a which is now of
composition Xi, ie it contains much less B than did the original a (compo-
sition X).
 atoms of metal 1A'
distribution

atoms of metal B'

of 'B%

composition X (fig.918)
distribution
of 'B'

crystal boundary
distribution
of1B"

X,(fig.9l8)

Fig. 9.19 The formation of coherent and non-coherent precipitates.

In (i) the alloy has been heated to Tu (Fig. 9.18) and then cooled rapidly (water quenched)
The structure consists of homogeneous solid solution a, supersaturated with B. (ii) Heating
to some selected temperature causes B atoms to migrate and form clusters within the a
lattice. These clusters produce |3\ which although preliminary to the formation of p, is still
continuous—or coherent—with the original a lattice, (iii) Here equilibrium has been attained
—the alloy has been heated to a sufficiently high temperature (9.91) so that |3 has been
rejected from a as a non-coherent precipitate. The lattice parameters of (3 no longer 'match
up' sufficiently well to those of a as do those of the intermediate P'. Hence a crystal boundary
now separates a and p.
 9.92 Precipitation under Non-equilibrium Conditions We will now
assume that the alloy X, which has been retained at temperature Tu, long
enough for its structure to be completely homogeneous a, is cooled very
rapidly to room temperature (TR). This could be achieved by quenching it
in cold water. In most cases treatment such as this will prevent any precipi-
tation from taking place and we are left with a solid solution which, at TR,
is now super-saturated with B. The structure will not be in equilibrium and
is said to be in a metastable state. It has an urge to return to a state of
equilibrium by precipitating some |3. At room temperature such precipi-
tation will be unlikely to occur due to the extreme sluggishness of move-
ment of B atoms, but if the temperature is increased, diffusion will begin
and then accelerate as the temperature continues to rise.
The alloy is held at some selected temperature so that diffusion occurs
at a low but definite rate. The temperature chosen is well below Ts in
order to ensure that non-coherent precipitation does not occur. During this
heat-treatment clusters of atoms of both A and B, but with the overall
composition of the phase |3, slowly form groups at many points in the a
lattice. These clusters have the important property that their lattice structures
are continuous with that of the a matrix, and there is no discontinuous
interface as exists between a and |3, which has precipitated during cooling
under equilibrium conditions. This unbroken continuity of the two lattices
is known as coherency. Since the cluster size is extremely small and the
rate of diffusion very slow, a large number of these coherent nuclei will
form and the chosen temperature will not be high enough to allow the
formation of a separate (3 structure. Instead, this intermediate structure—
we will call it (3'—is produced, and the mismatching between (3' and the a
matrix leads to distortion in the a lattice in the neighbourhood of these
nuclei. Such distortions will hinder the movement of dislocations and so
the strength and hardness increase. The greater the number of nuclei and
the larger they are, provided that coherency is retained, the greater the
strength and hardness of the structure as a whole. Time and temperature
will influence this 'precipitation' procedure and, hence, also the ultimate
mechanical properties which result. If the temperature is high (Ti, Fig.
9.20) a high rate of diffusion prevails, and this in turn leads to the formation
COHERENT
PRECIPITATION
TENSILE STRENGTH

NON-COHERENT
PRECIPITATION

TENSILE STRENGTH IF ALLOY IS SLOWLY COOLED

TO GIVE NON-COHERENT PRECIPITATE.

DURATION OF TREATMENT

Fig. 9.20
of relatively few nuclei, which, however, will grow to a large size. At a
lower temperature (TY) a larger number of nuclei will form and grow slowly
so that strength increases slowly.

Alloys Containing More Than Two Metals

9.100 In this chapter we have dealt only with binary alloys, that is, alloys
containing two different metals. In the case of ternary alloys, namely, those
containing three different metals, a third variable is obviously introduced,
so that the system can no longer be represented by a two-dimensional
diagram. Instead we must work in three dimensions in order adequately to
represent the alloy system. The base of the three-dimensional diagram will
be an equilateral triangle, each pure metal being represented by an apex
of the triangle. Temperature will be represented by an ordinate perpen-
dicular to the triangular base. Fig. 9.21 illustrates a ternary-alloy system
of the metals A, B and C in which the binary systems A-B, B-C and A-C
are all of the simple eutectic type. The three binary liquidus lines now
become liquidus surfaces in the three-dimensional system, and these sur-
faces intersect in three 'valleys' which drain down to a point of minimum
temperature, vertically above EABC- This is the ternary eutectic point of
the system, its temperature being lower than that of any of the binary
eutectic points EAB, EBC or EAc
Equilibrium diagrams for alloys of four or more metals cannot be rep-
resented in a single diagram since more than three dimensions would be
required. In practice such a system can often usefully be represented in
the form of a pseudo-binary diagram in which the concentration of one
component is varied whilst the others are kept constant—this is equivalent
to taking a vertical 'slice' through a ternary system, parallel to one of the
end faces. One of these pseudo-binary systems is shown in Fig. 14.1. Such
systems can be interpreted in the same way as an ordinary binary system.

Fig. 9.21 Diagram representing a three-dimensional ternary system, of three metals A, B

and C.
The curved isothermal contour lines meet binary eutectics which converge to a ternary
eutectic point, EABC-
 We have dealt here only with equilibrium diagrams of the simple funda-
mental types. Many of those which the reader will encounter are far more
complicated and often contain such a multitude of different phases as
almost to exhaust the Greek alphabet. This is particularly true of equilib-
rium diagrams which represent those copper-base alloy systems in which
a number of intermetallic compounds are formed and in which peritectic
reactions are also common. In general, we are interested only in those
parts of the diagram near to one end of the system, where we usually find
a solid solution with, possibly, small amounts of an intermetallic com-
pound. The interpretation then becomes much simpler, and the reader has
been provided with sufficient information in this chapter to deal with most
of the alloy systems likely to be encountered.

Rapid Solidification Processes (RSP)

9.110 It was explained earlier (3.18) that the crystal size present in a
casting is dependent upon the rate at which the metal is cooling when it
reaches the solidification temperature. Large castings cool at a slow rate
of the order of l°C/s and will contain crystals of average grain diameter in
the region of 5 mm. However with very rapid cooling rates of approxi-
mately 104oC/s the crystal size will be no more than 10~3 mm, whilst if the
cooling rate is increased further to 106oC/s a completely amorphous struc-
ture can be retained with some alloys. That is, the structure is non-
crystalline and similar to that of glass.
9.111 The refinement—or elimination—of grain attendant upon RSP
may also produce the following structural changes:
1 The effects of minor segregation (3.41) are reduced since the same
amount of impurity is spread over a vastly increased area of grain boun-
dary so that its effect is 'diluted'. In amorphous structures segregation
will be virtually eliminated since grain boundaries are absent. In either
case greater homogeneity is achieved.
2 Rapid cooling can greatly extend the limits of solid solubility by introduc-
ing increased supersaturation. This can enhance precipitation hardening.
3 Some phases may be retained at ambient temperatures which could not
be obtained by orthodox quenching methods.
Reduction in crystal size and the consequent reduction in minor segre-
gation result in much higher strength, Young's modulus and in resistance
to corrosion (21.70); whilst magnetic properties can be improved in suitable
alloys. With some aluminium alloys the amounts of iron, manganese,
cobalt and nickel can be increased because more of these elements will be
retained in solid solution, due to rapid solidification, and so increase
strength. Improved properties of rapidly solidified superalloys of both
nickel and chromium as well as of titanium alloys are reported; whilst
amorphous alloys of iron, silicon and boron are very suitable for use in
transformer cores since they show an extremely low remanence (14.35)
and hysteresis loss.
 9.112 The practical difficulties involved in attaining very high rates of
cooling limit rapid solidification products to either very fine powder or thin
foil. Nevertheless methods of achieving the high rates necessary to produce
extremely small crystals or, ultimately, an amorphous structure, are ingeni-
ous and varied.
Smooth spherical powders in a wide range of alloys can be made by
pouring the molten alloy through a small-diameter refractory nozzle.
Below the nozzle jets of a suitable inert gas impinge on the metal stream
'atomising' it and scattering it as small droplets. These are rapidly cooled
by the gas and collected as a fine powder at the base of the 'atomisation
tower'.
In spray-deposition processes a 'gas-atomised' stream of metal droplets
is directed on to a cooled metal surface. Fig. 9.22(i) indicates the basis of
a method by which strip may be formed; whilst in the Osprey process the
sprayed droplets build up on a cooled rotating former to produce discs,
rings or billets for subsequent hot working. Spray-deposition has been used
to produce high-speed steel forms up to 1.5 tonnes in mass.
Ribbon or foil can be manufactured direct from the melt by a variety of
'melt-spinning' processes the principles of which are indicated in Fig.
9.22(ii). Here molten metal is fed on to the surface of a continuously
cooled rotating wheel, to produce thin foil which subsequently parts from
the wheel as contraction occurs. Thin foil in iron-neodymium-boron
alloys is produced by this method and then consolidated for the manufac-
ture of permanent magnets.
Pulsed laser techniques giving cooling rates as high as 1012oC/s have been
developed recently and will increase the scope of this work.

MELT MELT

gas atomised
spray

WHEEL WHEEL

Fig. 9.22 Typical rapid solidification processes: (i) spray forming; (ii) a melt-spinning
process.

Exercises
1. Fig. 9.23 represents the magnesium-silicon thermal equilibrium diagram,
(i) Given that X is a chemical compound derive its formula. (Atomic masses:
Mg-24.3; Si-28.1);
(ii) Annotate the diagram completely;
 0
C
liquid

silicon (% wt.)

Fig. 9.23.

i
(iii) What phases can co-exist at E?
(iv) What will happen if the temperature of the alloy represented by E is
raised by a small amount?
(v) What are the compositions of the phases present in an alloy containing
80% Si-20% Mg overall, at a temperature of 11000C?
(vi) In what proportions by mass will the phases in (v) be present?
(vii) Describe, step by step, what happens as an alloy containing 50% of each
element cools slowly from 12000C to 7000C;
(viii) What will be the compositions of the phases present at 7000C?
(ix) In what proportions will the phases in (viii) be present?
2. Beryllium (m.pt. 1282°C) and silicon (m.pt. 1414°C) are completely soluble as
liquids but completely insoluble as solids. They form a eutectic at 10900C
containing 61% silicon.
Draw the thermal equilibrium diagram and explain, with the aid of sketches,
what happens when alloys containing (i) 10% silicon, (ii) 70% silicon, solidify
completely. (9.40)
3. Fig. 9.24 represents the bismuth-antimony thermal equilibrium diagram.
(a) Would you expect the diagram to be of this type? Give reasons.
(b) Consider an alloy containing 70% Sb-30% Bi which is cooling slowly:
(i) At what temperature will solidification begin?
(ii) What will be the composition of the first solid to form?
(iii) What will be the compositions of the phases present at 5000C?
(iv) In what proportions by mass will the phases in (iii) be present?
(v) What will be the composition of the last trace of liquid which solidifies?
(vi) At what temperature will solidification be complete?
(c) Explain what would happen if the alloy containing 70% Sb-30% Bi cooled

0
C
liquid

liquid
+ solid

solid

antimony (%>wt.)

Fig. 9.24.
 rapidly from 6000C to 3000C. Sketch the type of microstructure you would
expect in the solid alloy. (9.50)
4. Again referring to Fig. 9.24, at what temperature will a 60% Sb-40% Bi alloy
contain 25% by mass of liquid and 75% by mass of solid? What will be the
compositions of liquid and solid at this temperature?
5. Two metals A and B have melting points 6000C and 4500C respectively. The
following results indicate the temperatures associated with discontinuities in
the cooling curves of the alloys indicated:

%B 10 20 35 50 55 60 75 90 95
0
C 545 495 410 330 300 315 365 415 430
0
C 450 300 300 300 — 300 300 300 370

If the maximum and minimum percentage solubilities of the two metals are
20% B in A, 10% B in A, and 10% A in B, 5% A in B, sketch and label the
equilibrium diagram. (Assume solubility lines are straight.)
(i) At what temperature will an alloy containing 30% B begin to solidify?
(ii) Describe the cooling of an alloy containing 30% B and sketch typical
microstructures.
(iii) What proportions of a and (3 would you expect in the eutectic alloy at the
eutectic temperature and at 00C? (9.60)
6. Di*aw a thermal equilibrium diagram representing the system between two
metals, X and Y, given the following data:
(i) X melts at 10000C and Y at 8000C;
(ii) X is soluble in Y in the solid state to the extent of 10.0% at 7000C and
2.0% at 0 0 C;
(iii) Y is soluble in X in the solid state to the extent of 20.0% at 7000C and
8.0% at 0 0 C;
(iv) a eutectic is formed at 7000C containing 40.0% X and 60.0% Y.
Describe what happens when an alloy containing 70.0% X solidifies and
cools slowly to 0 0 C.
Sketch the microstructures of an alloy containing 15.0% Y (a) after it
has cooled slowly to 0 0 C; (b) after it has been heated for some time at
7000C and then water quenched. (9.60 and 9.92)
7. Fig. 9.25 shows part of the aluminium-magnesium thermal equilibrium
diagram.
(i) What is the maximum solid solubility of magnesium in aluminium?
(ii) Over what temperature range will an alloy containing 7% Mg exist as a
single solid phase?
(iii) At what temperature does an alloy containing 5% Mg begin to melt on
heating?
(iv) An alloy containing 16% Mg is at 5200C. What are the compositions of
the phases present?
(v) In what proportions will the phases in (iv) be present?
(vi) What is the percentage increase in solubility of magnesium in aluminium
in an alloy containing 12% Mg as the temperature rises slowly from 600C
to 3500C?
(vii) Sketch the structure of an alloy containing 10% Mg (a) slowly cooled
from 4500C; (b) water quenched from 4500C
(viii) Which will be the stronger in (vii)—(a) or (b)l (9.90)
 °c
liquid

oC + liquid

OC

CU + /3

magnesium (0ToWt.)

Fig. 9.25.

°c

liquid

tin (°/owt.)

Fig. 9.26.

8. Fig. 9.26 shows part of the silver-tin thermal equilibrium diagram. Consider an
alloy which contains 18% by mass of tin, cooling under equilibrium conditions:
(i) At what temperature does solidification begin?
(ii) What is the composition of the initial solid which forms?
(iii) At what temperature is solidification complete?
(iv) What is the composition of the last trace of liquid?
(v) What are the natures and compositions of the phases present at 7100C?
(vi) In what proportions by mass are these phases present at 7100C?
(vii) Outline, step by step, the structural changes which occur as the alloy cools
between 8500C and 6000C. Illustrate your answer with sketches of the
phase structures at each stage. (9.70)
9. The tin-indium phase diagram is shown in Fig. 9.27.
(i) Annotate it completely starting from the right-hand side of the diagram;
(ii) Indicate on the diagram (a) any eutectic points (b) any peritectic points;
(iii) What is the maximum solubility of tin in indium?
 0
C

Fig. 9.27.

10. Tungsten dissolves 5.0% platinum at 24600C to form a solid solution a. The
composition of the liquid in equilibrium with a at this temperature is 61.0%
tungsten-39.0% platinum. At 24600C a peritectic reaction takes place between
a and the remaining liquid forming a solid solution (3, which contains 63.0%
tungsten-37.0% platinum. At 17000C, a contains 4.8% platinum and P contains
37.1% platinum.
Draw the thermal equilibrium diagram between 35000C and 17000C and
explain what happens when liquids containing (i) 80.0% tungsten; (ii) 62.0%
tungsten solidify slowly. (M.pts.: Tungsten—34000C; Platinum—1773°C) (9.70)
11. Fig. 9.28 shows part of the gold-tin thermal equilibrium diagram.
(i) Derive formulae for the three compounds X, Y and Z (Atomic masses:
Au-197.0; Sn-118.7);

°c

liquid

gold Wowt.)

Fig. 9.28.
 (ii) Annotate the diagram;
(iii) Describe, step by step, the phase changes which occur in an alloy contain-
ing 75% Sn-25% Au as it cools slowly from 3500C to 00C. (9.70)

Bibliography
Cahn, R. W., Physical Metallurgy, North Holland, 1980.
Copper Development Association, Equilibrium Diagrams for Binary Copper
Alloys.
Elliot, R. P., Constitution of Binary Alloys (First Supplement), McGraw-Hill, 1965.
Ferguson, F. D. and Jones, T. K., The Phase Rule, Butterworths, 1958.
*Hansen, M., Constitution of Binary Alloys, McGraw-Hill, 1958.
Martin, J. W., Precipitation Hardening, Pergamon, 1968.
Martin, J. W., Micromechanisms in Particle-hardening Alloys, Cambridge Univer-
sity Press, 1979.
Massalski, T. B., Binary Alloy Phase Diagrams (VoIs. I and II), American Society
of Metals, 1989.
Rhines, R. N., Phase Diagrams in Metallurgy, McGraw-Hill, 1956.
Shunk, F. A., Constitution of Binary Alloys (Second Supplement), McGraw-Hill,
1969.
Smithells, C. J., Metals Reference Book, Butterworths, 1983.

* This work was first published in Berlin in 1936 as Der Aufbau der Zweistofflegierungen but is still a
major reference book for phase diagrams when used in conjunction with the Supplements by Elliot and
Shunk.