Page 287 - Electrical Properties of Materials

Soft magnetic materials 269

applications the best material is the one with the largest saturation magnet-
ization. As the frequency increases, it is still important to have large saturation
magnetization, but low coercivity is also a requirement. At high frequencies,
considering that eddy current losses are proportional to the square of the
frequency, the most important property is high resistivity.
Do losses matter? In practical terms this is probably the most important
materials science problem that we have touched upon. Something like many
millions of megawatts of electricity are being generated around the world, all
by generators with hysteresis losses of order 0.5–1.0%. Then a large fraction
of this electricity goes into motors and transformers with more iron losses.
If all inventors were paid a 1% royalty on what they saved the community,
then a good way to become rich would be to make a minute improvement
to magnetic materials. Is there any good scientific way to set about this? Not
really. We know that anisotropy, magnetostriction, and local stresses are bad,
but we cannot start from first principles and suggest alloys which will have
the required properties. The considerable advances that have been made in
magnetic materials have largely been achieved by extensive and expensive trial
and error. To Gilbert’s seventeenth-century crack about ‘good luck’ we must
add diligence for the modern smelters of iron. The currently used phrase is
actually ‘enlightened empiricism’.
Iron containing silicon is used in most electrical machinery. An alloy with B(T) (a)
about 2% silicon, a pinch of sulfur, and critical cold rolling and annealing pro- 1.5
cesses is used for much rotating machinery. Silicon increases the resistivity, 1.0
which is a good thing because it reduces eddy-current losses. Iron with a higher
0.5
silicon content is even better and can be used in transformer laminations, but
0.0
it is mechanically brittle and therefore no good for rotating machinery. Where
small quantities of very low-loss material are required and expense is not im- –0.5
portant, as for radio-frequency transformers, Permalloy [78.5% Ni, 21.5% Fe] –1.0
is often used. A further improvement is achieved in the material called ‘Super- –1.5
–0.5 0 0.5
malloy’ which contains a little molybdenum and manganese as well. It is very H(A / m )
–1
easily magnetizable in small fields [Fig. 11.9(a)] and has no magnetostriction.
B(T) (b)
We may now mention a fairly new and rather obvious trick. If anisotropy
1.5
is bad, and anisotropy is due to crystal structure, then we should get rid of Alnico 9
1.0
the crystal structure. What we obviously need is an amorphous material. How Alnico
can we produce an amorphous material? We can produce it by cooling the 0.5 5
melt rapidly, so that the liquid state disorder is frozen in. The key word is 0.0
‘rapidly’. In fact, the whole process is called Rapid Solidification Technology, –0.5
abbreviated as RST. The cooling should proceed at a speed of about a mil- –1.0
lion degrees per second, so the technological problems have not been trivial.
–1.5
In the first successful commercial solution a stream of molten metal is squir- –15 –10 –5 0 5 10 15 10 4
–1
ted on a cooled rotating drum, followed usually by a stress relief anneal at H(A / m )
◦
about 300 C. The resulting magnetic material has the form of long thin rib-
Fig. 11.9
bons typically about 50 μm thick and a few millimetres wide. New production
Hysteresis loops of (a) Supermalloy
methods, for example the planar flow casting method, in which a stable rect-
and (b) Alnico 5 and 9. Note the
angular melt ‘puddle’ feeds the material into the drum, have led to further factor 10 between the horizontal
5
improvements. It is now possible to obtain a uniform ribbon with a thickness scales of (a) and (b).
of 20–30 μm and up to 20 cm wide. The main advantage of amorphous mater-
ials is that they can be produced easily and relatively cheaply, with magnetic
properties nearly as good as those of commercial alloys, which require careful