Permanent Magnet Materials:

the most common classes.

by Dr. Peter Campbell

In Section 1B, we derived the ideal intrinsic demagnetization curve for a material which possesses magnetocrystalline anisotropy, with which a fully magnetized permanent magnet maintains a uniform magnetization of +M_sat until a reverse applied field equal to its intrinsic coercivity -H_ci flips the magnetization over to -M_sat. Magnetocrystalline anisotropy occurs to some degree in all types of permanent magnet, and exclusively in many. Section 1C concluded by relating the second quadrant actual B vs. H and intrinsic demagnetization curves to their ideal counterparts, so we can now introduce the most important classes of permanent magnet material and explain to what extent their demagnetization curves reflect the existence of magnetocrystalline (and/or another type of) anisotropy. More extensive details of the properties of these and all other classes of commercial permanent magnets are given in subsequent Sections; this Section introduces:

ceramic ferrite | samarium-cobalt | neodymium-iron-boron

Ceramic Ferrite

Many permanent magnets are made by powder metallurgical methods using powder that has been milled to fine particles. To maximize the saturation magnetization M_sat in the bulk material, it is clearly advantageous if the magnetic moments within each particle are aligned, and the net magnetizations from all the particles are aligned. The latter is achieved by applying an orienting field during the process of consolidating the powder into a solid magnet. The moments within each particle will be spontaneously aligned as shown in diagram (a) provided that the powder is milled to about the size of a single magnetic domain. If the particle size is much larger than this, it will be energetically more favorable for a domain wall to exist as shown in diagram (b), setting up no net magnetization from the particle. It is possible to calculate the critical diameter below which the powder must be milled (based upon energy considerations), which is approximately 1µm in the case of ceramic ferrite magnets.

Ceramic ferrites are made using an iron oxide powder, to which either barium or strontium is added to improve alignment of the crystal lattice structure. The formula is M.n(Fe₂O₃), where M = Ba or Sr and 5.8<n<6.0. After milling, the powder is pressed in a die, with an orienting field applied through the cavity if desired. If no field is applied, then an isotropic magnet will result with no preferred magnetic properties in any particular direction; if an orienting field is applied, then an anisotropic magnet will result having preferred magnetic properties along a given axis. The compacted powder is then sintered at a temperature of 1100-1300^oC (hence the name "ceramic") to achieve full densification, and lastly ground to its final dimensions. Alternatively, the powder may be blended with a polymer binder, and then either extruded or formed in a die by compression- or injection-molding, thus producing a bonded ferrite magnet of near net shape; again, anisotropic properties may be achieved by applying an orienting field through the die cavity.

Because ceramic ferrites use single domain particles, they base their permanent magnetism on magnetocrystalline anisotropy, with characteristics closely following those described in Sections 1B and 1C. As an example, the diagram below shows the demagnetization characteristics measured at different temperatures for a "Ceramic 8" grade of ceramic ferrite magnet. The intrinsic demagnetization curves (shown dashed in this diagram) and the "normal" B vs. H curves (shown solid) compare well with the actual characteristics for a magnet exhibiting magnetocrystalline anisotropy that were illustrated in Section 1C.

ceramic 8

Samarium-cobalt

Atoms of the rare earth elements tend to form intermetallic compounds with transition metals such as Fe, Ni or Co, and in the early development of "rare earth" magnets, theory predicted that the "light" rare earth elements (particularly samarium) would combine most favorably with cobalt to produce a high crystal anisotropy. Processing of rare earth-cobalt predicts the feasibility of several intermetallic compounds, the first practical magnet being made from the compound SmCo₅, followed later by Sm₂Co₁₇. Samarium-cobalt magnets are manufactured by much the same routes as ferrites, being formed either into a fully dense sintered magnet, or a compression- or injection-molded bonded magnet. Because both samarium and cobalt are relatively expensive elements, anisotropic magnets are usually produced with optimized properties along a pre-determined axis.

Rare-earth magnets have a rather more complicated domain wall mechanism than that for pure magnetocrystalline anisotropy, such that the best magnetic properties are achieved with the powder milled only to a grain size which is an order of magnitude larger than the single domain size. This means that, not only can domain walls exist, but they can move relatively freely within a grain. While this allows the saturation magnetization to be achieved with only a modest applied field, a high intrinsic coercivity will depend upon the grains' ability to resist the formation of a reverse domain when a demagnetizing field is applied. This vital property is therefore controlled by the grain boundaries, which are composed of deviations from the primary composition that provides a strong pinning of the domain walls at these sites. This mechanism, known as nucleation, occurs in SmCo₅ magnets; a grain undergoing nucleation is shown in diagram (a) below.

Sm₂Co₁₇ differs from SmCo₅ in that its grains contain a fine cell structure as illustrated here in diagram (b). Heat treatment of this compound promotes the formation of these Sm₂Co₁₇ cells, separated by thin walls of SmCo₅ which now provide pinning of the domain walls (rather than the grain boundaries). Pinning, rather than nucleation, is therefore the controlling mechanism in Sm₂Co₁₇ magnets, and this requires that a much greater field be applied to initially magnetize it to saturation.

Whether it is a nucleation-type SmCo₅ or a pinning-type Sm₂Co₁₇ magnet, i.e. whether the domain walls are pinned at the grain or cell boundaries, they will move quite freely once these pinning forces are overcome, and M_sat will flip over into the opposite direction quite abruptly when an applied field of -H_ci is reached. This is the same behavior that was described in the derivation of intrinsic coercivity for magnetocrystalline anisotropy, and so we would expect to see similarly shaped demagnetization curves for samarium-cobalt type magnets, which we do in the diagrams below. From the discussion on "Change in Coercivity" in Section 3A, the material's intrinsic coercivity is proportional to its crystallographic constant K₁, which incorporates its magnetic dipole moment µ_m; the change in µ_m, K₁ and hence H_ci with temperature is clearly more severe in diagram (b): Sm₂Co₁₇ than it is in (a): SmCo₅.

samarium cobalt

Neodymium-iron-boron

After the successful development of samarium-cobalt magnets, there was concern about the availability and cost of these two principal elements. Fe is a much cheaper transition metal than Co, and Nd is a considerably more plentiful "light" rare earth element than Sm. Several rare earth elements (R) were combined with Fe in R₂Fe₁₇ compounds, but all were found to have very low potential operating temperatures. A significant improvement was then made with the discovery that the addition of boron yielded a ternary compound with strong uniaxial magnetocrystalline anisotropy, and a much improved operating temperature. A neodymium-iron-boron compound which approximated to "Nd₂Fe₁₄B" was found to offer the best combination of magnetic and thermal properties. Today's commercial neodymium-iron-boron magnets have many combinations of partial substitutions for Nd and Fe, leading to a wide range of available properties.

Sintered:

There are several different routes by which neodymium-iron-boron magnets are manufactured. One method is similar to that for ceramic ferrite and sintered samarium-cobalt, the milled powder being formed into a fully dense anisotropic magnet by compaction in an orienting field and then sintering. This produces a nucleation-type magnet as described and illustrated in diagram (a) above, in which the grain boundaries are composed of deviations from the primary Nd₂Fe₁₄B composition, providing pinning of the domain walls. One of the problems with this conventional route to producing neodymium-iron-boron is that the powder is highly susceptible to oxidation at the grain boundaries, which seriously limits the particle size to which it can be milled, and really makes it impractical for use in bonded magnets.

Rapidly quenched:

A completely different process route involves rapid quenching of the molten neodymium-iron-boron alloy, using a "melt spinning" technique to produce a ribbon which is then milled to powder. While the crushed ribbon yields relatively large platelet-shaped powder particles, rapid quenching provides them with the extremely fine microstructure illustrated here in diagram (c), again having grain boundaries which deviate from the primary Nd₂Fe₁₄B composition. This is not, however, a nucleation-type magnet, because the powder has such a fine microstructure that it conforms to the single domain model described above for ceramic ferrite. Neodymium-iron-boron magnets that are made from rapidly quenched powder base their permanent magnetism on magnetocrystalline anisotropy, and require a much greater applied field to initially align the grains' magnetizations and magnetize this material to its saturation level. As diagram (c) suggests, it is not practical to mill this powder to single domain size, so rapidly quenched powder is inherently isotropic. However, it can be consolidated into a fully dense anisotropic magnet by the plastic deformation which occurs in hot pressing. The natural protection afforded the grain boundaries by the fine microstructure also makes this powder very stable against oxidation, so it is easy to blend and form into any type of isotropic bonded magnet.

HDDR:

We have mentioned the problem of oxidation in the preparation of neodymium-iron-boron powder, so it is not surprising to learn that it also readily absorbs hydrogen, which turns the material into a very brittle powder. This may actually be used to advantage to make the powder more amenable to milling, and has become the basis of the "HDDR" process (which stands for hydrogenation, disproportionation, desorption and recombination, obviously too complex to worry about the details here!). The HDDR process also gives neodymium-iron-boron powder an ultrafine structure with grains about the size of a single domain, and the powder can be milled into particles of around this size. HDDR prepared neodymium-iron-boron powder is therefore inherently anisotropic, and magnets made from it base their permanent magnetism on magnetocrystalline anisotropy. HDDR powder can be hot pressed into a fully dense anisotropic magnet, or it can be blended and molded into an anisotropic bonded magnet.

In summary, neodymium-iron-boron magnets may be made from powder which is:

sintered, nucleation-type;
rapidly quenched, magnetocrystalline;
HDDR, magnetocrystalline.

Neodymium-iron-boron magnets therefore exhibit similarly shaped demagnetization characteristics to the other classes described above, with well-defined "knees" at which M_sat reverses as an applied field of -H_ci is approached. As an example, the diagram below shows the demagnetization characteristics measured at different temperatures for one grade of fully dense anisotropic Nd₂Fe₁₄B .