Ferromagnetic Domains

What makes iron ferromagnetic

Every electron is a tiny magnet, carrying a spin magnetic moment of about one Bohr magneton, μ_B = 9.274 × 10⁻²⁴ A·m². In most materials these moments either pair up and cancel or point in random directions, so the material is at most weakly paramagnetic or diamagnetic. In iron, nickel and cobalt something stronger happens: a quantum-mechanical exchange interaction — a consequence of the Pauli exclusion principle and Coulomb repulsion, not a magnetic force — energetically favors neighboring spins being parallel.

The exchange energy between two neighboring spins is captured by the Heisenberg model:

E_exchange = −2 J S_i · S_j

where J is the exchange integral. When J > 0 (ferromagnets) the energy is lowest when the spins are parallel, so over a region every moment locks into the same direction. This is why a region of iron is spontaneously magnetized all the way to saturation even with no external field. The catch is that aligning all of the iron in one direction would create an enormous external magnetic field, and that costs energy. Nature's compromise is to break the material into domains.

Why domains form

A single-domain block of iron stores a large amount of energy in the magnetic field outside it — the magnetostatic or stray-field energy. By splitting into two oppositely-magnetized domains, the external field is largely confined, roughly halving that energy. Splitting again into four domains halves it again, and adding triangular closure domains at the ends can almost completely cancel the external field. Pierre Weiss first postulated these regions in 1907; we now call them Weiss domains.

So domains are the result of an energy balance between several competing terms:

Energy term	Wants to…	Origin
Exchange energy	Keep neighboring spins parallel	Pauli + Coulomb (quantum)
Magnetostatic (stray field) energy	Break into many domains, close the flux	Self-field of the magnet
Magnetocrystalline anisotropy	Point moments along "easy" crystal axes	Spin–orbit coupling to lattice
Domain-wall energy	Minimize total wall area	Exchange + anisotropy in the wall
Magnetoelastic (magnetostriction) energy	Match strain in the lattice	Coupling of M to crystal strain

The actual domain pattern — slab domains, closure domains, fir-tree or maze patterns — is whatever configuration minimizes the sum of these energies for a given sample shape and crystal orientation.

Domain walls

Between two domains the magnetization cannot flip abruptly — that would cost too much exchange energy. Instead it rotates gradually over a thin transition layer called a domain wall. In a bulk material the rotation is usually a Bloch wall (the moment rotates out of the plane of the wall); in thin films it is often a Néel wall (rotation stays in plane). The wall width δ and energy per unit area σ result from a tug-of-war:

δ ≈ π √(A / K)        σ ≈ 4 √(A K)

where A is the exchange stiffness (≈ 2 × 10⁻¹¹ J/m for iron) and K is the magnetocrystalline anisotropy constant (≈ 4.8 × 10⁴ J/m³ for iron). Exchange wants a thick, gradual wall; anisotropy wants a thin one. The result for iron is a wall about 40–100 nm wide containing a few hundred atomic planes. A 180° wall separates anti-parallel domains; a 90° wall separates perpendicular domains common in closure structures.

The magnetization process: how a field magnetizes iron

When you ramp up an external field H, the sample magnetizes in three distinct regimes, all visible as moving boundaries in the visualization above:

1. Reversible wall motion (weak field). Domains already pointing close to H grow slightly; walls bow and shift but snap back if you remove the field. Magnetization rises gently and reversibly.
2. Irreversible wall motion (moderate field). Walls break free of pinning sites — inclusions, grain boundaries, dislocations — and jump forward in discrete steps. These are Barkhausen jumps; you can literally hear them as crackling noise in a pickup coil. Most of the magnetization is gained here, and it does not reverse on its own.
3. Rotation toward saturation (strong field). Once the favorable domains have eaten the others, the only domains left point along an easy axis that is off from H. The strongest fields rotate this magnetization out of its easy axis into the field direction. When every moment is aligned, the sample is saturated — magnetization can rise no further.

At saturation the magnetization equals M_s, the saturation magnetization. For iron M_s ≈ 1.71 × 10⁶ A/m, corresponding to a saturation flux density μ₀M_s ≈ 2.15 T. The atomic moment of iron works out to about 2.2 μ_B per atom.

Hysteresis and magnetic memory

Because the irreversible Barkhausen jumps cannot undo themselves, the magnetization does not retrace its path when H is reduced — the M–H curve forms a closed loop, the hysteresis loop. Two numbers describe it:

• Remanence M_r — the magnetization left when H returns to zero. This is what makes a permanent magnet permanent and what stores a bit on a hard-disk platter.
• Coercivity H_c — the reverse field needed to bring the magnetization back to zero. It measures how strongly the walls are pinned.

The area enclosed by the loop is the energy dissipated as heat per unit volume per cycle. Transformer cores cycle 50 or 60 times a second, so this loss matters enormously — which is why core makers use "soft" materials with skinny loops. The contrast between soft and hard materials is entirely a contrast in how easily domain walls move:

Property	Soft magnet (Si-steel, permalloy)	Hard magnet (NdFeB, SmCo, ferrite)
Coercivity H_c	~1–100 A/m	~10⁵–2 × 10⁶ A/m
Wall motion	Easy, low pinning	Strongly pinned / single-domain particles
Hysteresis loop	Tall and narrow, low loss	Wide, high stored energy
Energy product (BH)_max	Negligible	Up to ~400 kJ/m³ (NdFeB)
Typical use	Transformer & motor cores, inductors	Permanent magnets, motors, speakers

The Curie temperature: switching ferromagnetism off

The exchange interaction aligns spins; thermal agitation randomizes them. As temperature rises, thermal energy k_B·T grows until it overwhelms exchange. Above the Curie temperature T_C the ordered domains disappear and the material becomes paramagnetic, its susceptibility following the Curie–Weiss law:

χ = C / (T − T_C)      (for T > T_C)

Material	Curie temperature T_C	μ₀M_s at room temperature
Iron (Fe)	770 °C (1043 K)	2.15 T
Cobalt (Co)	1115 °C (1388 K)	1.79 T
Nickel (Ni)	358 °C (631 K)	0.61 T
Gadolinium (Gd)	20 °C (293 K)	2.6 T (below T_C)
Neodymium magnet (Nd₂Fe₁₄B)	312 °C (585 K)	1.6 T

This is also a practical demagnetizing trick: heat a magnet past T_C and it forgets its magnetization. Conversely, cooling iron from above T_C in zero field re-forms a random multi-domain state — fully demagnetized.

How we know domains are real

• Bitter patterns. Sprinkle a colloid of fine magnetic particles on a polished surface; they collect at domain walls where the stray field is strongest, drawing the walls out under a microscope (Francis Bitter, 1931).
• Magneto-optical Kerr effect (MOKE). Polarized light reflecting off the surface rotates differently for up- vs down-domains, imaging domains optically in real time.
• Barkhausen noise. A coil around the sample picks up voltage spikes as walls jump — direct evidence that magnetization changes in discrete, irreversible steps.
• Lorentz microscopy and electron holography. Electrons deflected by the in-sample field map domains and walls at nanometre resolution.

Where ferromagnetic domains show up

• Data storage. Each bit on a hard disk or magnetic tape is a small region whose remanent magnetization direction encodes 0 or 1; shrinking the bit means controlling ever-smaller domains.
• Transformers and motors. Soft cores rely on easy wall motion so that magnetization follows the alternating current with minimal hysteresis loss.
• Permanent magnets. Motors, generators, loudspeakers, MRI machines and hard-drive actuators all use hard materials whose pinned walls give large coercivity and remanence.
• Magnetic shielding. High-permeability mu-metal guides flux around sensitive instruments because its walls move with almost no field.
• Sensors. Fluxgate sensors, magnetostrictive actuators and Barkhausen-noise probes (for non-destructive testing of steel stress) all exploit domain behavior.
• Spintronics. Racetrack memory and domain-wall logic deliberately push individual walls along nanowires with spin-polarized currents.

Common misconceptions

• "Domains are individual atoms." No — a single domain holds 10¹²–10¹⁸ atoms. The atoms within one domain are already perfectly aligned by exchange.
• "Magnetizing iron creates magnetism from nothing." The domains were already saturated; the field only reorganizes which way they point, by moving walls and rotating moments.
• "The exchange interaction is a magnetic force between spins." It is electrostatic plus Pauli exclusion in origin; the dipole–dipole magnetic force between atoms is hundreds of times too weak to explain ferromagnetism.
• "You can keep magnetizing iron indefinitely." Once saturated, all moments point along H and magnetization cannot increase — only the flux density B keeps rising as μ₀H.
• "Soft and hard refer to mechanical hardness." They refer to magnetic coercivity. Some soft magnets are physically hard alloys, and vice versa.
• "Heating always strengthens a magnet." Heating toward and past the Curie temperature weakens and ultimately destroys the magnetization.