the partial substitution of Co for Fe.
When neodymium-iron-boron was first introduced, a
major advantage was seen to be its freedom from the availability and cost
issues that periodically plague the use of cobalt. But as noted in the
previous Section 3C, the partial substitution of Co for Fe in
Nd2Fe14B
does have the advantage of reducing surface
oxidation. This, together with the extremely fine microstructure of
rapidly quenched neodymium-iron-boron powder, yields a bonded
Neo magnet with inherently good corrosion resistance. A more
complete description of the role that Co plays follows.
The Curie Temperature Tc of
Co is much greater than that of Fe, so when Co is progressively substituted for
Fe in Nd2(Fe1-xCox)14B, Tc
rises above the
310oC
of the basic ternary alloy as shown in the diagram below left, at
an approximate rate of
+10oC
per atomic % of Co. This increase in Tc
leads to an improvement in the reversible temperature coefficient of Br (α),
the magnitude of α being
seen to decrease in the diagram below right. However, as this diagram
also shows, with regard to intrinsic coercivity there is a worsening of the reversible temperature coefficient of
Hci (β) with Co
content, which is due to a combination of factors as described below.
Rapidly quenched powder for bonded magnets
As
Co progressively substitutes for Fe, the
saturation magnetization Msat
increases slightly to a maximum value and decreases thereafter. This
causes a similar effect on the remanence Br,
as shown here on the right for an isotropic Nd2(Fe1-xCox)14B
powder that is made by rapid quenching. In Section 1B
we derived the relationship between
intrinsic coercivity Hci and Msat
as
but then we explained in Section 3A that K1
would not be constant in materials such as neodymium-iron-boron.
In the case of Nd2(Fe1-xCox)14B,
the graph to the left here (Y. Matsuura et al, "Magnetic
properties of the Nd2(Fe1-xCox)14B
system", Appl. Phys. Lett., 46(3), 1 Feb 1985, pp.
308-310) shows how K1 falls
as the proportion of Co (x) is increased. The net effect is for Hci
to decline (as shown above for rapidly quenched powder) and for its
reversible temperature coefficient β
to worsen with Co content. A more precipitous decline in Hci
(akin to the discussion in Section 3B) worsens linearity of the
demagnetization curves and leads to poorer irreversible flux loss. Nevertheless, when combining the magnetic
properties as in the determination of maximum energy product, it is clear
that (BH)max
increases, at least up to a certain percentage of Co in a Nd2(Fe1-xCox)14B
type alloy.
In practice, the best compromise between all of these
effects is achieved with Co substituted at a relatively low level, as
demonstrated below for rapidly quenched powders with zero and 2
atomic % of Co, and for compression-bonded magnets made from these powders
(2 atomic % is approximately equivalent to x=0.025 in Nd2(Fe1-xCox)14B).
The powders and magnets underwent a quite common corrosion test in
an atmosphere of 85% relative humidity at a temperature of 85oC. The inset photographs show
the results after 500 hours' exposure, with rusting much more apparent in
both the powder and
the magnets with no Co. A more quantitative assessment is made by
measuring weight gain (%) in the samples, for which the powder with 2 atomic
% Co is clearly superior (lower weight gain) to that with no Co. This
advantage is carried over into the compression-bonded Neo magnets, the
greater weight gain in both types being attributable to absorption in the
magnets' epoxy binder.
In summary, a small amount of Co (substituted for Fe) in
powder for bonded Neo magnets provides beneficial effects:
- much better corrosion resistance;
- slightly higher Br and
(BH)max;
- smaller α.
But this also causes undesirable effects:
- worse irreversible flux loss;
- higher β;
- greater alloy cost.
Copyright © 2005 by Princeton
Electro-Technology, Inc.
All rights reserved.
