Ferroelectricity: BaTiO3, Domains, and the Curie Point

From an Ordinary Dielectric to a Ferroelectric

In guide 2 you met the ordinary dielectric and its polarization: switch on an electric field and its bound charges shuffle a little, storing energy; switch the field off and the shuffle relaxes straight back to zero. A ferroelectric breaks that rule. It carries a built-in spontaneous polarization — a permanent separation of plus and minus charge inside every unit cell that survives with no field at all — and, crucially, an applied field can switch that polarization to point the other way. Picture an ordinary dielectric as a room of weathervanes that swing only while the wind blows, and a ferroelectric as a room of little compass needles that stay pointed on their own and that you can deliberately flip end-for-end.

Why care? A ferroelectric's dielectric constant is enormous — a few thousand, against roughly 10 for a plain oxide insulator — and that giant permittivity is exactly what lets guide 2's multilayer ceramic capacitor store so much charge in a chip the size of a sesame seed. The workhorse ferroelectric is barium titanate, BaTiO3. So this is no laboratory curiosity: it is the physics behind the most-manufactured electronic component on Earth — and, as guide 4 will show, behind piezoelectric sensors and actuators too.

Inside Barium Titanate: A Marble in One of Two Dimples

BaTiO3 is a perovskite, the ABO3 family: big Ba2+ ions at the eight corners of a cube, O2- ions at the six face centres forming an octahedral cage, and one small Ti4+ ion sitting inside that oxygen octahedron. Above the Curie temperature the cell is perfectly cubic and centrosymmetric — the Ti4+ sits dead centre, the centre of positive charge falls exactly on the centre of negative charge, and there is no dipole. This hot state is paraelectric: an ordinary (if unusually high-permittivity) dielectric, nothing more.

Cool below the Curie point and the cage distorts. The cube stretches a little along one axis into a tetragonal cell (the c/a ratio grows to about 1.01 — a 1 percent elongation), and the Ti4+ slides roughly 0.1 Angstrom (about 0.01 nm) off-centre toward one of the oxygens. Picture a marble that has rolled into one of two shallow dimples in the floor of its cage. Now the positive and negative charge centres no longer coincide, so every unit cell is a permanent electric dipole. Add up all those aligned dipoles and the crystal carries a spontaneous polarization of about 0.26 C/m^2 — a genuinely large number.

Why exactly two dimples? The off-centre spot is a lower-energy state than dead-centre, and there are two equivalent ones — up or down the tetragonal axis — so the marble must pick one. This is a double-well: two stable minima with an energy hump between them, and 'switching' means kicking the marble over that hump with a field (the origin of the hysteresis loop below). The shift is also cooperative: neighbouring cells lower their energy by pointing the same way, so whole regions polarize together instead of each cell choosing at random. That cooperation is what makes the polarization macroscopic and switchable — the true fingerprint of ferroelectricity. And because the charges are now displaced, the structure is non-centrosymmetric, which (guide 4) is the strict requirement for piezoelectricity.

The Curie Point: Where Ferroelectricity Switches Off

Heat BaTiO3 and at its Curie temperature Tc — about 120 to 130 degrees C — the tetragonal distortion vanishes: the marble climbs back to dead centre, the cell snaps to cubic, the spontaneous polarization disappears, and the crystal becomes an ordinary paraelectric. Cool back down and it all reappears. The Curie point is the on/off temperature of ferroelectricity, exactly as it is for ferromagnetism in iron. And right at Tc something dramatic happens to the permittivity: it does not fade — it spikes, climbing to a sharp peak of tens of thousands, because the lattice is balanced on a knife-edge between cubic and tetragonal and so is softly, easily polarized.

1. Above about 130 degrees C: cubic and centrosymmetric — paraelectric, no spontaneous polarization.
2. About 5 to 130 degrees C (room temperature lives here): tetragonal — the Ti4+ shifts along a cube edge. This is BaTiO3's everyday ferroelectric state.
3. About -90 to 5 degrees C: orthorhombic — the polarization axis swings round to point along a face diagonal.
4. Below about -90 degrees C: rhombohedral — the polarization points along a body diagonal of the cube. Each step is just a further tilt of the same off-centre idea, and each shows up as a bump in the permittivity-versus-temperature curve.

Above Tc the permittivity obeys the Curie-Weiss law, eps ~ C/(T - Tc): the closer you sit above the Curie point, the higher it climbs. That is both a gift and a curse for the capacitor maker. A gift, because engineers park the working temperature just below a Curie peak to harvest a permittivity of thousands; a curse, because pure BaTiO3's capacitance swings wildly around 130 degrees C, useless for a stable part. The fix, hinted at in guide 2, is to dope and layer the BaTiO3 — shifting Tc with additives and building core-shell grains — to smear the sharp peak into a broad, flat plateau, trading peak height for temperature stability to meet a spec like X7R.

Domains and the Hysteresis Loop

If every cell in a fresh BaTiO3 crystal is a dipole, why doesn't a freshly fired capacitor cling to your finger like a charged comb? Because the crystal breaks itself into domains — patches within which all the dipoles point the same way, but which point in different directions from patch to patch, so the whole thing cancels to nearly zero net polarization. It does this to escape the huge electrostatic energy of one giant dipole (all that uncompensated surface charge) and to relieve the mechanical strain of the tetragonal stretch. It is the exact electric echo of why a fresh iron nail is not a magnet even though every atom in it is one: the domains are scrambled. The thin walls between them — 180-degree walls where the polarization simply reverses, and 90-degree walls where it turns a corner and the strain flips too — are where the switching action happens.

Apply a field and the favourably-oriented domains grow at the expense of the rest: the domain walls sweep sideways, swallowing badly-aligned patches, until the whole crystal points one way. This domain-wall motion is the real, microscopic reason the permittivity is so large and so nonlinear — you are not merely stretching bonds, you are marching walls across the crystal. Push far enough and it saturates, every domain aligned. Do this deliberately with a strong DC field and lock it in, and you have poled the ceramic — the step (guide 4) that turns a randomly-oriented ferroelectric into a working piezoelectric.

          P (polarization)
          ^
    +Ps --|------------.--o     (1) ramp field E up: domains
          |        _.-'             align, P saturates at +Ps
    +Pr --o----_.-'             (2) E back to 0: P stays = +Pr
          |   /                     (remanence -- the "memory")
   -------+--o-----------> E
          | /|  +Ec
    -Pr --o' |                  (3) reverse E to -Ec: P = 0,
          |  |                      the dipoles flip over
    -Ps --o--'                  (4) push to -Ps, then loop back:
          |                         the return path is a fat LOOP

Trace polarization P against field E as you cycle up and back and you do not retrace your path — you sweep out a fat loop, the hysteresis loop, the very signature of a ferroelectric. Read three numbers off it. The saturation polarization Ps, where every domain is aligned; the remanent polarization Pr, what remains when you remove the field (this is the memory); and the coercive field Ec — about 1 kV/cm for bulk BaTiO3 — the reverse field you must apply to drag P back through zero and flip it. The loop's fatness is not free: its enclosed area is energy dumped as heat every cycle, the ferroelectric share of the dielectric loss you met in guide 2, which is why these high-permittivity dielectrics run lossier than a humble paraelectric. Warm the crystal above Tc and the loop collapses to a thin straight line — no remanence, no switching, ferroelectricity gone.

Why One Small Distortion Powers Electroceramics

Stand back and marvel at how much rides on a titanium ion sitting one-tenth of an Angstrom off-centre. That single non-centrosymmetric shift hands BaTiO3 the giant permittivity behind the multilayer ceramic capacitor — trillions made every year — and makes the crystal switchable, the basis of ferroelectric memory. Because a poled, non-centrosymmetric ferroelectric couples electricity to shape, the very same physics delivers PZT's piezoelectricity (guide 4); and because the spontaneous polarization itself changes with temperature, it delivers the pyroelectric infrared detectors of guide 5. One tiny distortion, an entire discipline.

From here, everything is a variation on the theme. Guide 4 poles a ferroelectric to make a piezoelectric, then builds sonar, medical ultrasound, and spark igniters from it. Guide 5 puts the same doped BaTiO3 to work differently — its grain boundaries plus this Curie transition give the self-regulating PTC thermistor — and reads its temperature-dependent polarization as a pyroelectric sensor. One marble, two dimples, and a temperature at which it climbs back to the middle: that is the seed of half of electroceramics.