A phenomenon observed in certain solids in which the magnetic properties change abruptly at a certain characteristic temperature known as the Curie point. Below the Curie point the solid exhibits ferromagnetic properties. Above this temperature the thermal energy of the atoms is sufficient to produce magnetic properties typical of paramagnetism: the susceptibility obeys the Curie-Weiss law approximately, the value of the Weiss constant being close to that of the Curie point and a few degrees higher.

The chief ferromagnetic elements are iron, cobalt, and nickel; there are also many ferromagnetic alloys based on these materials. Ferromagnetic materials are characterized by a large positive susceptibility: very large values of magnetization are produced by relatively small magnetic fields, the magnetization varying nonlinearly with field strength (Fig. a). Maximum intensity of magnetization (magnetic saturation) is achieved at fairly low field strengths and a certain amount of magnetization is retained when the magnetizing field is removed, hence the materials exhibit magnetic hysteresis.

(a) Typical magnetization curve of a virgin ferromagnetic specimen

Ferromagnetism was first explained by Weiss, who suggested that spontaneous magnetization occurs within ferromagnetic materials due to large interatomic forces acting between neighbouring atoms in the crystal lattice. Below the Curie point these forces can overcome the thermal effects and tend to produce an ordered state. The interatomic forces were discovered by Heisenberg and are known as exchange forces. Weiss also postulated that the groups of atoms are organized into tiny bounded regions called domains. In each individual domain the magnetic moments of the atoms are aligned in the same direction. The domain is thus magnetically saturated and behaves like a magnet with its own magnetic moment and axis. In an unmagnetized sample the domains are randomly orientated so that the magnetization of the specimen as a whole is zero. The existence of domains has been verified experimentally.

The magnetic moments of the atoms arise from the spin of electrons in an unfilled inner shell. In any stable material in the absence of an applied magnetic field, the detailed arrangements of the magnetic moments result from the interaction between the various forces operating within the sample. It is always such as to produce the minimum energy possible. In ferromagnetic materials this minimum energy state occurs when the electron spins of the atoms within a domain are arranged in parallel.

Magnetization of the domains is much harder along certain directions relative to the crystal axes than in others: more energy is required to magnetize the domains lying along these directions. This anisotropy energy is least for small domains. To form the boundaries between the domains however also requires energy because of the exchange forces between neighbouring atoms and this tends to increase domain size. The domain size is determined by a compromise between these competing forces. The boundaries are called Bloch walls and extend over a finite number of atoms, each of whose spins are slightly displaced relative to that of its neighbours (Fig. b). The energy state is also affected by the degree of crystalline perfection, the existence of strains and impurities affecting significantly the ferromagnetic behaviour. It can be shown that in any particular material the energy state of the domains is least when the Bloch wall intersects as many dislocations as possible. A typical energy curve as a function of wall position is shown in Fig. c. A virgin specimen would have a wall located at the minimum energy position, marked A.

(b) Representation of spin direction in a Bloch wall with 180˚ rotation (c) Energy as a function of Bloch wall position

When a magnetic field is applied to a ferromagnetic material the characteristic shape of the magnetization curve (Fig. a) is explained by consideration of the domain behaviour. At small values of magnetic field the net effect is to displace the Bloch walls over a few atoms, away from the minimum energy state; thus those domains with spins parallel or nearly parallel to the field grow at the expense of the others (Fig. d). If the field is removed the walls tend to move back to the minimum energy state and for small values of applied field the magnetization changes are small and reversible. At larger values of applied field the wall excursions are sufficiently large so that an energy maximum, marked B on Fig. c, is passed through and the change becomes irreversible. A single crystal with few dislocations allows much greater reversible wall excursions than a polycrystalline material with many strains and impurities present; it also requires a lower field to produce them. As the magnetic field is increased further a position is reached when further domain growth becomes impossible; further magnetization is only possible by rotation of the magnetic axes of the domains. This is a more difficult process than domain growth because of the crystal anisotropy and above the knee of the magnetization curve the magnetization increases only slowly until saturation is reached.

(d) Representation of magnetization of a ferromagnetic sample

Ferromagnetic materials are classified as either hard or soft. Hard materials have a low relative permeability, very high coercive force, and are difficult to magnetize and demagnetize; soft materials have a high relative permeability, low coercive force, and are easily magnetized and demagnetized.

Hard ferromagnetics, such as cobalt steel and various ferromagnetic alloys of nickel, aluminium, and cobalt, retain a high percentage of their magnetization and have a relatively high hysteresis loss (see magnetic hysteresis). They are most suitable for use as permanent magnets, as used in loudspeakers. A high degree of dislocation is introduced into their structure during manufacture. Hard materials are frequently heated to high temperatures and then quenched in a suitable liquid to introduce strains. Alternatively they may be produced as compressed powders in which each particle is sufficiently small so as to be a single domain; magnetization can then only proceed by domain rotation since the energy required for wall movement is so great.

Soft ferromagnetics, such as silicon steel and soft iron, retain very little magnetization and have extremely small hysteresis loss. The ease of magnetization and demagnetization makes them very suitable for uses involving changing magnetic flux, as in electromagnets, electric motors, generators, and transformers. They are also useful for magnetic screening. Their properties are enhanced by careful manufacture, as by heating and slow annealing, in order to achieve a high degree of crystal purity.

Although the large magnetic moment at room temperatures makes soft ferromagnetic materials extremely useful for magnetic circuits, most ferromagnetics are very good conductors and suffer energy loss from eddy currents produced within them. The ideal material for magnetic circuits would be a ferromagnetic insulator. There is also an additional energy loss due to the fact that magnetization does not proceed smoothly but in minute jumps (see Barkhausen effect). This loss is known as magnetic residual loss and depends purely on the frequency of the changing flux density, not on its magnitude. See also antiferromagnetism.