microscopical examination gives valuable information concerning
the suitability' of a sample of steel for special purposes.
Mixture by fusion is the general method of producing an
alloy, but it is not the only method possible. It would
seem, indeed, that any process by which the particles of
two metals are intimately mingled and brought into close
contact, so that diffusion of one metal into the other can
take place, is likely to result in the formation, of an
alloy. For example, if vapours of the volatile metals
cadmium, zinc and magnesium are allowed to act on platinum or
palladium, alloys are produced. The methods of manufacture
of steel by cementation, case-hardening and the Harvey
process are important operations which appear to depend on
the diffusion of the carburetting material into the solid
metal. When a solution of silver nitrate is poured on to
metallic mercury, the mercury replaces the silver in the
solution, forming nitrate of mercury, and the silver is
precipitated; it does not, however, appear as pure metallic
silver, but in the form of crystalline needles of an alloy
of silver and mercury. F. B. Mylius and O. Fromm have shown
that alloys may be precipitated from dilute solutions by zinc
cadmium, tin, lead and copper. Thus a strip of zinc plunged
into a solution of silver sulphate, containing not more than
0.03 gramme of silver in the litre, becomes covered with a
flocculent precipitate which is a true alloy of silver and
zinc, and in the same way, when copper is precipitated
from its sulphate by zinc, the alloy formed is brass. They
have also formed in this way certain alloys or definite
composition, such as AuCd3, Cu2Cd, and. more interesting
still, Cu2Sn. A very similar fact, that brass may be formed
by electrodeposition from a solution containing zinc and
copper, has long been known. W. V. Spring has shown that by
compressing a finely divided mixture of 15 parts of bismuth,
8 parts of lead, 4 parts of tin and 3 parts of cadmium, an
alloy is produced which melts at 100 deg. C., that is. much
below the melting-point of any of the four metals. But these
methods or forming alloys, although they suggest questions
of great interest, cannot receive further discussion bore.
Our knowledge of the nature of solid alloys has been much
enlarged by a careful study of the process of solidification.
Let us suppose that a molten mixture of two substances A and
B, which at a sufficiently high temperature form a uniform
liquid, and which do not combine to form definite compounds,
is slowly cooled until it becomes wholly solid. The phenomena
which succeed each other are then very similar, whether A
and B are two metals, such as lead and tin or silver and
copper, or are a pair of fused salts, or are water and common
salt. All these mixtures when solidified may fairly be termed
alloys.1 If a mixture of A and B be melted and then allowed
to cool, a thermometer immersed in the mixture will indicate
a gradually falling temperature. But when solidification
commences. the thermometer will cease to fall, it may even
rise slightly, and the temperature will remain almost constant
for a short time. This halt in the cooling, due to the heat
evolved in the solidification of the first crystals that form
in the liquid is called the freezing-point of the mixture;
the freezing-point can generally be observed with considerable
accuracy. In the case of a pure substance, and of a certain
small class of mixtures, there is no further fall in temperature
until the substance has become completely solid, but, in
the case of most mixtures, after the freezing-point has been
reached the temperature soon begins to fall again, and as the
amount of solid increases the temperature becomes lower and
lower. There may be other halts in the cooling, both before
and after complete solidification, due to evolution of
heat in the mixture. These halts in temperature that occur
during the cooling of a mixture should be carefully noted,
as they give valuable information concerning the physical and
chemical changes that are taking place. If we determine the
freezing-points of a number of mixtures varying in composition
from pure A to pure B, we can plot the freezing-point
curve. In such a curve the percentage composition can be
plotted horizontally and the temperature of the freezing-point
vertically, as in fig. 5. In such a diagram, a point P defines
a particular mixture, both as to percentage, composition and
temperature; a vertical line through P corresponds to the
mixture at all possible temperatures, the point Q being its
freezing-point. In the case of two substances which neither
form compounds nor dissolve each other in the solid state,
the complete freezing-point curve takes the form shown in
fig. 5. It consists of two branches AC and BC, which meet
in a lowest point C. It will be seen that as we increase
the percentage of B from nothing up to that of the mixture
C, the freezing-point becomes lower and lower, but that if
we further increase the percentage of B in the mixture, the
freezing-point rises. This agrees with the well-known fact
that the presence of an impurity in a substance depresses its
melting-point. The mixture C has a lower freezing or melting
point than that of any other mixture; it is called the eutectic
mixture. All the mixtures whose composition lies between
that of A and C deposit crystals of pure A when they begin to
solidify, while mixtures between C and B in composition deposit
crystals of pure B. Let us consider a little more closely the
solidification of the mixture represented by the vertical line
PCRS. As it cools from P to Q the mixture remains wholly
liquid, but when the temperature Q is reached there is a
halt in the cooling, due to the formation of crystals of A.
The cooling soon recommences and these crystals continue to
form, but at lower and lower temperatures because the still
liquid part is becoming richer in B. This process goes on
until the state of the remaining liquid is represented by
the point C. Now crystals of B begin to form, simultaneously
with the A crystals, and the composition of the remaining
liquid does not alter as the solidification progresses.
Consequently the temoerature does not change and there is
another well-marked halt in the cooling, and this halt lasts
until the mixture has become wholly solid. The corresponding
changes in the case of the mixture TUVW are easily understood
--the first halt at U, due to the crystallization of pure
B, will probably occur at a different temperature, but the
second halt, due to the simultaneous crystallization of A and
B, will always occur at the same temperature whatever the
composition of the mixture. It is evident that every mixture
except the eutectic mixture C will have two halts in its
cooling, and that its solidification will take place in two
stages. Moreover, the three solids S, D and W will differ
in minute structure and therefore, probably, in mechanical
properties. All mixtures whose temperature lies above the
line ACB are wholly liquid, hence this line is often called
the ``liquidus''; all mixtures at temperatures below that
of the horizontal line through C are wholly solid, hence
this line is sometimes called the ``solidus,'' but in more
complex cases the solidus is often curved. At temperatures
between the solidus and the liquidus a mixture is partly solid
and partly liquid. This general case has been discussed at
length because a careful study of it will much facilitate the
comprehension or the similar but more complicated cases that
occur in the examination of alloys. A great many mixtures
of metals have been examined in the above-mentioned way.
Fig. 6 gives the freezing-point diagram for alloys of lead and
tin. We see in it exactly the features described above.
The two sloping lines cutting at the eutectic point are the
freezing-point curves of alloys that, when they begin to
solidify, deposit crystals of lead and tin respectively. The
horizontal line through the eutectic point gives the second
halt in cooling, due to the simultaneous formation of lead
crystals and tin crystals. In the case of this pair of metals,
or indeed of any metallic alloy, we cannot see the crystals
forming, nor can we easily filter them off and examine them
apart from the liquid, although this has been done in a few
cases. But if we polish the solid alloys, etch them if
necessary, and examine them microscopically, we shall find
that alloys on the load side of the diagram consist of
comparatively large crystals of lead embedded in a minute
complex, which is due to the simultaneous crystallization
of the two metals during the solidification at the eutectic
temperature. If we examine alloys on the tin side we shall
find large crystals of tin embedded in the same complex. The
eutectic alloy itself, fig. 2 (Plate), shows the minute complex
of the tin-lead eutectic, photographed by J. A. Ewing and W.
Rosenhain, and fig. 3 (Plate), photographed by F. Osmond, shows
the structure of a silver-copper alloy containing considerably
more silver than the eutectic. Here, the large dark masses
are the silver or silver-rich substance that crystallized
above the eutectic temperature, and the more minute black
and white complex represents the eutectic. It is not safe
to assume that the two ingredients we see are pure silver
and pure copper; on the contrary, there is reason to think
that the crystals of silver contain some copper uniformly
diffused through them, and vice versa. It is, however, not
possible to detect the copper in the silver by means of the
microscope. This uniform distribution of a solid substance
throughout the mass of another, so as to form a homogeneous
material, is called ``solid solution,'' and we may say that
solid silver can dissolve copper. Solid solutions are probably
very common in alloys, so that when an alloy of two metals
shows two constituents under the microscope it is never safe
to infer, without further evidence, that these are the two
pure metals. Sometimes the whole alloy is a uniform solid
solution. This is the case with the copper-tin alloys
containing less than 9% by weight of tin; a microscopic
examination reveals only one material, a copper-like substance,
the tin having disappeared, being in solution in the copper.
Much information as to the nature of an alloy can be obtained
by placing several small ingots of the same alloy in a furnace
which is above the melting-point of the alloy, and allowing
the temperature to fall slowly and uniformly. We then extract
one ingot after another at successively lower temperatures
and chill each ingot by dropping it into water or by some
other method of very rapid cooling. The chilling stereotypes
the structure existing in the ingot at the moment it was
withdrawn from the furnace, and we can afterwards study this
structure by means of the microscope. We thus learn that
the bronzes referred to above, although chemically uniform
when solid, are not so when they begin to solidify, but that
the liquid deposits crystals richer in copper than itself,
and therefore that the residual liquid becomes richer in
tin. Consequently, as the final solid is uniform, the
crystals formed at first must change in composition at a later
stage. We learn also that solid solutions which exist at
high temperatures often break up into two materials as they
cool; for example, the bronze of fig. 1, which in that figure
shows two materials so plainly, if chilled at a somewhat
higher temperature but when it was already solid, is found
to consist o
