Polymorphism as a Change of Structure

One element, many crystals: what polymorphism is

Every rung so far has treated a crystal as one fixed thing: a lattice with a motif stamped at every point, the same wallpaper repeating forever. But here is a fact that should feel a little uncanny. Take a single element — say iron — hold its chemistry perfectly constant, and it can still freeze into more than one distinct crystal structure. Which structure you get depends only on temperature and pressure. That capacity is called polymorphism (for a compound) or allotropy (for an element), and it is the doorway to this whole rung. We will read every transformation not as a heat-treatment recipe but as a change of STRUCTURE — the wallpaper being re-stamped with a different unit cell while the ink stays exactly the same.

Iron is the poster child. Below 912 degrees C, iron is alpha-iron, a body-centred cubic (BCC) crystal — a cube with an atom at each corner and one in the middle. Heat it past 912 degrees C and, still solid, it re-stamps itself into gamma-iron, a face-centred cubic (FCC) crystal, atoms at the corners and one on each face. Push past 1394 degrees C and it flips back to BCC (delta-iron) until it finally melts at 1538 degrees C. Same iron atoms throughout; three different arrangements, swapped by temperature alone. The whole of steelmaking rests on this single trick — that iron can be coaxed to switch its crystal structure without ever melting.

Why the structure switches: a contest of free energy

Why would iron ever bother to rebuild itself? Because at a given temperature and pressure, a material adopts whichever structure has the lowest Gibbs free energy, G = H - T times S. Here H is the enthalpy (roughly, how tightly the atoms are bonded — lower is more stable) and S is the entropy (roughly, how much disorder and vibrational freedom the structure allows — higher is more stable). Enthalpy usually favours the tightest-bound, often densest packing. Entropy favours the structure whose atoms can jiggle most freely. The two pull in different directions, and the T times S term is the referee: because it grows with temperature, the higher-entropy structure inevitably wins once T is high enough.

Two slopes make this precise and let you predict the direction of any transition. First, dG/dT = -S: the higher-entropy phase has a steeper downward G(T) curve, so as you heat, its curve dives below the other and it takes over — that crossing point is the transition temperature. Second, dG/dP = V: raise the pressure and the smaller-volume (denser) phase is favoured. Those two facts explain a startling amount at a glance — BCC iron winning at high temperature (its soft, low-frequency vibrations carry extra entropy), and diamond winning at high pressure (it is denser than graphite). One honest caveat: the folk rule high temperature equals a more open structure has exceptions; the true driver is entropy, mostly vibrational, which usually but not always tracks openness.

Let us put real numbers on iron's switch, because it hides a surprise. BCC packs to an atomic packing factor of 0.68 with coordination 8; FCC packs tighter, 0.74 with coordination 12. So gamma (FCC) is the denser packing. Working it out: alpha-iron has a = 2.87 angstrom with 2 atoms per cell, about 11.8 angstrom^3 per atom; gamma-iron extrapolates to a = 3.57 angstrom with 4 atoms per cell, about 11.4 angstrom^3 per atom — roughly 3 percent denser. At the real 912 degrees C the measured jump is about a 1 percent CONTRACTION. Read that again: iron gets SMALLER as you heat it through 912 degrees C. (The atomic radius is not quite identical in the two phases, so the clean 0.68-versus-0.74 argument is a simplification — but the direction is right, and this contraction is exactly why quenching steel builds up internal stresses.)

Three vivid switches: iron, tin, and carbon

Tin gives the most theatrical example, and it runs the other way in temperature. Ordinary metallic tin — white tin, beta-tin, a body-centred tetragonal metal at 7.27 g/cm^3 — is only stable above about 13.2 degrees C. Below that, tin WANTS to become grey tin (alpha-tin), which has the wide-open diamond-cubic structure at just 5.77 g/cm^3. That is a huge roughly 26 percent volume EXPANSION, so the metal swells, cracks, and crumbles into grey powder — the disease called tin pest, sped up by deep cold. It has historically wrecked organ pipes and reputedly the tin buttons on soldiers' coats. Notice the daughter phase is the very same structure as diamond and silicon.

Carbon carries the deepest lesson: which phase is STABLE is a different question from which phase you actually GET. At room temperature and 1 atmosphere, graphite (layered, 2.26 g/cm^3) is the thermodynamically stable form, and diamond (diamond-cubic, 3.51 g/cm^3) is only metastable — its free energy is very slightly higher. Yet a diamond sits on a ring for a million years without ever reverting, because turning back into graphite means breaking every single covalent bond and rebuilding — a colossal energy barrier that room temperature cannot climb. Diamond is thermodynamically doomed but kinetically immortal. And to make diamond in the first place you need high PRESSURE (dG/dP = V favours the denser diamond) together with high temperature (to give atoms enough mobility to clear that barrier) — exactly the mantle-like conditions of a synthesis press.

THREE ALLOTROPES, THREE SWITCHES   (APF = atomic packing factor)

  ELEMENT   ONE FORM                    OTHER FORM                  SWITCH & EFFECT
  -------------------------------------------------------------------------------------
  iron      alpha  BCC   APF 0.68       gamma  FCC   APF 0.74       heat past 912 C:
            coord 8                      coord 12  (denser)          ~1% CONTRACTION
  tin       beta   white, tetragonal    alpha  grey, diamond        cool below 13 C:
            metallic  7.27 g/cm3         cubic     5.77 g/cm3        ~26% EXPANSION -> crumbles
  carbon    graphite  layered           diamond   diamond-cubic     graphite stable at 1 atm;
            2.26 g/cm3  (stable)         3.51 g/cm3  (metastable)    diamond needs high P + T

How the atoms actually move: reconstructive vs displacive

So far we have only asked which structure wins. The other half of this rung asks HOW the atoms get there — the mechanism — and it splits into two great families. A reconstructive transformation is a wholesale rebuild: bonds break, atoms become mobile and diffuse, and they re-sort into the new packing. It is slow, needs thermal activation to clear a large barrier, and can move a lot of atoms a long way. Graphite-to-diamond is the extreme case — no wonder it needs a press and a furnace. Guide 2 of this rung is devoted to exactly this: the reconstructive way, contrasted with its opposite.

1. A nucleus of the new phase forms, and at its edge the old bonds must break.
2. Atoms freed from their old sites gain mobility and diffuse — this is the step that demands heat and takes time.
3. They re-sort and settle onto the sites of the new structure, in its new coordination.
4. The new-phase region grows as this disordered boundary sweeps forward through the old crystal.

The opposite family is the displacive transformation: no bonds break and no atom diffuses. Instead every atom shifts only a small fraction of a bond length, all of them together, cooperatively — the whole block simply shears into the new structure like a deck of cards skewed sideways. Because nothing has to diffuse, it can happen almost instantly and even at very low temperature. Its archetype is the martensitic transformation, the diffusionless shear that hardens steel — the star of guide 3, with its crisp orientation relationship, its habit plane, its shape change, and its transformation twinning. A third relative, covered in guide 4, is the order-disorder transformation: the atoms barely move at all, but their chemical SORTING changes, as a random solid solution re-sorts onto separate sublattices to become an ordered superlattice like Cu3Au.

Symmetry, order parameters, and domains: a map of this rung

There is one lens that unifies almost every structural transition: symmetry. Cooling from the high-temperature parent phase to the low-temperature daughter almost always LOWERS the symmetry — the daughter keeps only some of the parent's symmetry operations. Crystallographers say the daughter's space group is a subgroup of the parent's, a group-subgroup relation. This has a beautiful and very physical consequence. Because the parent had symmetry operations the daughter lacks, those lost operations relate several equally valid ways the daughter can form. Where two such variants meet, you get a domain wall — the twin boundaries, antiphase boundaries, and ferroelastic domains you will meet again and again. Symmetry lowering is, quite literally, what breeds domains.

To measure the change we use an order parameter — a single number that is zero in the symmetric parent and grows in the daughter. If it JUMPS discontinuously, the transition is first-order: it has a latent heat, a sudden volume change, and hysteresis (iron, tin, and most reconstructive switches are first-order). If it grows CONTINUOUSLY from zero, the transition is second-order (continuous), described beautifully by Landau theory, and often driven by a soft mode — a lattice vibration whose frequency drops toward zero as the atoms lean into the very displacement that freezes in the new structure. The ferroelastic and ferroelectric transitions are exactly this kind, and they are the subject of guide 5.