How a micron-long coil becomes a crystal
Guide 1 taught you that a free chain in the melt is a random coil, a floppy ball whose size is captured by its radius of gyration; guide 2 taught you that a stereoregular chain — an isotactic or syndiotactic one, with the right tacticity — has a clean repeating shape and so, in principle, could pack into a lattice. Now put the two facts together and you hit a genuine puzzle. A modest polyethylene chain of molar mass 100,000 g/mol is about 3,600 ethylene units long; stretched straight it would run roughly 1 micrometre, yet in the melt it is balled up into a coil only tens of nanometres across. If such enormous, tangled chains crystallize, what on earth does the crystal look like? The naive guess — chains lined up fully extended, like a bundle of matches — essentially never happens on its own.
The answer, discovered around 1957 (Keller, Fischer, and Till, working independently), is startling. Polyethylene crystallized from dilute solution grows as thin, flat, lozenge-shaped single-crystal plates only about 10 nm (100 angstrom) thick — yet electron diffraction shows the chain axis pointing straight THROUGH that thin dimension, perpendicular to the flat face. The chains are a hundred times longer than the plate is thick, so there is only one possible resolution: the chain folds. It runs across the ~10 nm thickness, makes a hairpin turn at the surface, re-enters, and runs back — over and over. This is the chain-folded lamella, and a single micron-long chain folds back on itself something like a hundred times, like a fire hose flaked back and forth on a deck, or a long ribbon of dough folded into a loaf pan.
Inside the lamella, though, is a real three-dimensional crystal with true long-range order and a proper unit cell — for polyethylene it is orthorhombic, roughly a = 7.4, b = 4.9, c = 2.53 angstrom, with the short c repeat (one all-trans ethylene unit) running along the chain. Notice the bonding asymmetry this forces: strong covalent bonds run ALONG the chain, straight through the thin dimension of the plate, while only weak van der Waals bonds hold neighbouring chains together side to side. That is exactly why polymer crystals are soft, mechanically anisotropic, and thin. And here is the key kinetic fact: the lamella thickness is set by the crystallization temperature, NOT by the chain length — crystallize hotter and the folds get longer and the plates thicker; the chain length just decides how many lamellae one chain threads through.
The semicrystalline reality: never one hundred percent
Here is the honest headline that separates polymers from metals and salts: a bulk polymer crystallized from the melt is NEVER fully crystalline. It is semicrystalline — an intimate two-phase interleaving of crystalline lamellae and amorphous regions. Why can it not simply finish crystallizing? Because the chains are long and entangled. One chain threads through several lamellae and the disordered layers between them; chain ends, branches, and stray atactic defects get rejected from the tight lattice and pile up as amorphous material. A chain that leaves one lamella, wanders through the amorphous layer, and enters the next is a tie molecule — and those tie molecules, stitching crystal to crystal across the gaps, are precisely what give a semicrystalline plastic its toughness rather than the brittleness of a pure crystal.
SEMICRYSTALLINE POLYMER ( chain = |, hairpin fold = __, ~ = amorphous )
ONE CHAIN-FOLDED LAMELLA (~10 nm thick) STACK of lamellae + amorphous
__ __ __ __ ||||||||| lamella (crystal)
/ \ / \ / \ / \ <- fold surface ~~~~~~~~~~~ amorphous layer
| | | | | | | | chain axis c runs ||||||||| lamella
| | | | | | | | THROUGH the thin ~~~o~~~~~~ o = tie molecule
\__/ \__/ \__/ \__/ dimension ||||||||| lamella
covalent bonds ALONG c (through the thickness); van der Waals ACROSS
-> the crystal is thin, soft, and highly anisotropic
A SPHERULITE = lamellae radiating from ONE nucleus, grown to impingement:
\ | / seen under CROSSED POLARIZERS
\ | / a dark MALTESE CROSS appears:
----( o )---- nucleus +---------+
/ | \ | \ / |
/ | \ | \ / |
(amorphous polymer trapped | / \ |
between the radiating ribs) | / \ |
+---------+Because it is two phases of different density, you can weigh a polymer's crystallinity straight off its bulk density. Chains pack tighter in the crystal — higher coordination, optimized van der Waals contacts — so crystalline polyethylene (rho_c about 1.00 g/cm^3) is denser than amorphous polyethylene (rho_a about 0.85 g/cm^3). The two-phase percent crystallinity follows from a simple mixing rule: percent crystalline = [rho_c x (rho_s - rho_a)] / [rho_s x (rho_c - rho_a)] x 100. Work a case: a high-density polyethylene of measured density rho_s = 0.97 g/cm^3 gives [1.00 x (0.97 - 0.85)] / [0.97 x (1.00 - 0.85)] x 100 = 0.12 / 0.1455 x 100, about 82 percent crystalline. A branched low-density polyethylene at 0.92 g/cm^3 comes out near 51 percent — the branches literally cannot fit the lattice and are exiled to the amorphous phase. Differential scanning calorimetry (heat of fusion) and wide-angle X-ray diffraction (sharp Bragg peaks riding on a broad amorphous halo) give the same fraction by independent routes.
One honest caveat before we move on: the clean two-phase model is itself a simplification. In reality there is a third, interfacial region — the rigid amorphous fraction — where chains emerging from the fold surfaces are neither fully ordered nor fully free to move. That is why density, calorimetry, and diffraction rarely agree to the last percent: each quietly defines crystalline a little differently. So percent crystallinity is a genuinely useful, reproducible-enough number, but not a fundamental constant of the material — always report how you measured it.
The spherulite: lamellae that radiate
In dilute solution you get those isolated flat lamellae. But crystallize from the melt — the industrially normal case — and something more dramatic unfolds. A crystal nucleates at a point, and lamellae grow radially outward in every direction, branching and splaying as they go, until they sweep out a roughly spherical aggregate called a spherulite, anywhere from a micrometre to a millimetre across. Between the radiating crystalline ribs sits the amorphous polymer the advancing crystals rejected. So the spherulite is itself two-phase all over again: crystalline lamellae as the radiating spokes, amorphous polymer filling the gaps, tie molecules bridging.
- A nucleus forms — a chance ordered cluster of chain segments, or a seed on a deliberately added nucleating-agent particle.
- A lamella grows outward as a folded ribbon, its chain axis lying tangential, that is, perpendicular to the growth radius.
- The ribbon splays and branches non-crystallographically, fanning out from a slender sheaf into an ever-widening cone.
- Branching fills every angle until the aggregate closes into a sphere: a spherulite radiating from its central nucleus.
- Neighbouring spherulites expand until they collide and impinge, meeting at flat boundaries that tile space — just as grains in a metal fill space by growth to impingement.
Slide a thin polymer film between crossed polarizers and each spherulite reveals itself as a striking dark Maltese cross, its four dark brushes aligned with the polarizer and analyzer directions, often crossed by concentric rings (banded spherulites, from lamellae that slowly twist as they grow). Be honest about what you are seeing: the Maltese cross is the optical fingerprint of radial, birefringent order — extinction wherever the lamellae align with the crossed axes — not a direct picture of the lamellae themselves. Spherulite size is a lever on properties, a clean structure-property relationship: coarse spherulites scatter light (a hazy film) and crack easily along their weak amorphous boundaries (brittle), while dusting in a nucleating agent creates many tiny spherulites at once, giving a clearer, tougher plastic. It is the polymer-processing echo of grain refinement from the polycrystal rung.
Two temperatures: melting and the glass transition
A fully crystalline solid has one thermal landmark, a melting point; a fully amorphous polymer has one too, a glass transition. A semicrystalline polymer, being both at once, has BOTH. It has a glass transition temperature Tg, where its amorphous fraction unfreezes from a rigid glass into mobile rubber, and a melting temperature Tm, where its crystalline lamellae finally melt. Always Tg is below Tm. Between the two, the material is a fascinating self-made composite: rubbery, mobile amorphous regions reinforced by hard crystalline lamellae. That is precisely the state of a polyethylene grocery bag at room temperature — its Tg sits far below (around -120 degrees Celsius), so the amorphous part is soft and rubbery, while its crystals (Tm about 135 degrees Celsius) stay solid and lend the film its strength and stretch.
One subtlety separates polymer melting from a metal's. A metal melts at one sharp temperature; a polymer melts over a broad RANGE, because its lamellae come in a spread of thicknesses and a thin lamella melts lower than a thick one. This is the Gibbs-Thomson effect — a crystal's melting point drops as it gets thinner, because its disordered fold surfaces cost energy that eats into the crystal's stability, the very same melting-point depression that shrinks the melting point of any tiny crystal. So the Tm you read off a chart is only an approximate peak; the true equilibrium melting point, of an infinitely thick perfect crystal, is higher than anything you ever measure. The glass transition, meanwhile, is the same free-volume freezing you met in the amorphous rung: as the amorphous chains cool, their free volume shrinks until chain segments can no longer rearrange in the time you are giving them, and the amorphous fraction locks in place.
Where semicrystalline order sits, and what comes next
Step back and place this on the order-disorder map you have been building all the way up this ladder. A metal or a salt is essentially fully crystalline — long-range order everywhere. A glass is fully amorphous — sharp short-range order but no long-range periodicity. A semicrystalline polymer is the great in-between: genuine three-dimensional crystals (the lamellae, with real long-range order and a proper unit cell) intimately interleaved with amorphous coils in the amorphous state (short-range order only), the two knitted together by tie molecules threaded across the fold surfaces. It is, in effect, a nanocomposite the material quietly makes of itself, blending the stiffness of the crystalline state with the give of the amorphous one.
But crystalline-versus-amorphous is not the only way a soft material can be partly ordered, and that is the thread the rest of this rung pulls. A liquid crystal keeps orientational order — all the rod-like molecules point the same way — while giving up positional order: order in direction without order in place, the subject of guide 4. And a block copolymer, or a soap, orders at a larger mesoscale entirely: chemically incompatible blocks phase-separate into regular lamellae, cylinders, and spheres tens of nanometres across through self-assembly, the subject of guide 5, where the very same folding chains you met here organize themselves without ever forming an atomic crystal. Semicrystalline structure is your first real taste of a soft material choosing partial order; the next two guides show you the other flavours it can choose.