June 6, 1912] 
NATURE sie) 
THE CRYSTALLISATION OF METALS? 
‘7 HE crystalline characters of metals have been 
much less completely studied than those of non- 
metallic minerals and artificial salts, owing in large 
part to the infrequency of occurrence of regular and 
Fic. 1.—Surface of galvanised iron. 
well-defined crystals amongst metals. Masses of 
metal are now known to be entirely crystalline, but 
special means are necessary in order to reveal their 
structure. In a few cases, notably that of bismuth, 
good results are obtained by pouring off the still 
liquid portion of a partly solidified mass of metal, 
when characteristic striated 
crystals of bismuth, recalling the 
Greek “key” pattern, result. ef 
Crystals are also obtained in relief hp 
on the surface of ingots cooled in 
contact with the air, tin, alu- f 
minium, and silver giving good 
results in this way. If the 
solidifying metal is spread out in 
a thin layer, the structure in relief 
may be developed in quite a re- 
markable degree, as when sheets 
of steel are dipped in molten zinc 
its 
in the preparation of ‘ galvan- 
ised"’ iron. The crystals (Fig. 1) 
closely resemble those of frost 
figures on glass. Crystals of steel 
up to 15 in. in length are occasion- 
ally found in the cavity or “ pipe”’ 
of large ingots, and these have a 
characteristic form—that of closely 
packed, spiky branches arranged 
at right angles to a main stem. 
The internal dendritic structure 
of a solid mass of crystalline metal 
is most readily revealed in the case 
of an alloy. By suitable etching, the primary crystal- 
lites may be brought into contrast with the material 
subsequently deposited. The arrangement of the 
axes of such crystal skeletons is not readily followed 
1 Abstract of a paver read before the Royal Philosophical Society of 
Glasgow on November 29, 1911, by Dr. Cecil H. Desch. 
NO, 2223, VOL. 89] 
<<): 
| 
| 
by the examination of the usual plane sections, and 
a better representation of the arrangement of the 
parts in space is obtained by adopting the biological 
method of serial sections. A specimen ts so ground 
as to present two accurately parallel faces, and is 
then placed, after etching, on the stage of the micro- 
scope in a special holder which 
permits the observer to bring 
the same area repeatedly before 
the objective. A suitable 
crystallite, having been selected, 
is photographed, and a thin 
layer is then removed from the 
surface by grinding and polish- 
ing. After again etching, the 
thickness is again measured, 
and a second photograph is 
taken. After several repeti- 
tions of this process the photo- 
graphs, which represent plane 
parallel sections of the speci- 
men, may be used for the re- 
construction of the crystallite 
in plastic material. In the 
specimen of phosphor-copper 
shown to the Society, fourteen 
such layers were removed, the 
average thickness of each 
layer being 0-014 mm. 
A marked feature of most 
metallic crystallites is the 
rounded termination of their 
axes. This rounding can only 
be attributed to the effects of 
surface tension at the moment 
of solidification. Intermetallic 
compounds are frequently less rounded, and less dis- 
posed to assume dendritic forms, than pure metals. 
In the varied patterns of eutectic alloys it is some- 
times difficult to recognise any relation to crystallisa- 
tion, and it is evident that surface-tension plays an 
essential part. In the copper-antimony alloy shown 
Fic. 2.—Eutectic of antimony and copper. 
in Fig. 2, however, it is seen that the minute anti- 
mony crystals of the eutectic are all in parallel 
orientation, and that the direction of their principal 
axes is the same as that of the neighbouring large 
crystallite. The violet copper antimonide forms a mere 
filling material, occupying the intervening spaces. 
