Why a polymer never fully crystallises
Back on the metals and ceramics rungs, crystallising was almost automatic: cool the melt and the atoms — little marbles — slot onto a lattice, building a crystalline solid with long-range order tiling to the horizon. A polymer cannot do that, and the reason is the very thing that makes it a polymer. Each macromolecule is a mile-long strand of thousands of repeat units, and in the melt these strands are tangled together like a bowl of cooked spaghetti. To crystallise perfectly, every strand would have to untangle and lie down in register with its neighbours — and there simply is not the time or the freedom for a knot that big to comb itself straight. So a polymer only ever crystallises in patches. It is semicrystalline: small ordered crystalline regions embedded in a sea of disordered, tangled amorphous chains.
How does even a patch manage it? Not by many separate chains lining up, but by a single chain folding back on itself over and over — imagine a fire hose flaked neatly back and forth into its hose bed, or a long ribbon folded accordion-style into a flat stack. Each straight run is only a few nanometres long, so only short segments need to register locally, which is achievable. The result is a thin, flat crystal platelet called a lamella, typically 10 to 20 nm thick, with the chain folds forming its top and bottom faces. The chain ends, the branches, the impurities, and all the hopeless tangles get shoved out into the amorphous regions between lamellae. The fraction of the material that ends up ordered this way is the percent crystallinity, and it ranges from essentially 0 up to about 95 percent — but never a clean 100.
Keep one subtlety close. A single chain is far longer than one lamella is thick, so one molecule can thread out of a crystalline lamella, wander through the amorphous region, and dive into the next lamella — a tie chain belonging to both worlds at once. Crystallinity in a polymer is therefore not an all-or-nothing property of each molecule; it is a statistic about how much of the material is locally ordered. Those tie chains matter enormously later: they are the stitches that carry load across the soft amorphous gaps, and losing them is often how a plastic goes brittle with age.
The spherulite: a polymer's version of a grain
Zoom out from a single lamella and you meet the polymer's answer to the metal grain. As a melt cools, lamellae do not grow in isolation; they sprout from a nucleus and fan outward in every direction, thin crystalline ribbons splaying like the vanes of a feather, with amorphous chains trapped in the gaps between them. The growing ball of radiating lamellae is a spherulite. It expands outward until it collides with its neighbours, and the spherulites jam together to fill all space — exactly the way metal grains grow until they impinge on a common grain boundary. A finished semicrystalline polymer is a mosaic of spherulites, each a micron to a millimetre across, and each one internally a blend of crystalline lamellae and amorphous filler.
You can actually see spherulites. Slice the polymer thin, put it between two crossed polarising filters under a microscope, and each spherulite lights up with a dark Maltese cross — the signature of lamellae radiating from a centre. That same radiating structure explains an everyday fact: spherulites are typically larger than the wavelength of light and scatter it, so a highly crystalline polymer looks cloudy or milky (think of an opaque HDPE milk jug or a translucent bucket), while a fully amorphous polymer with no spherulites to scatter light can be glass-clear (a polystyrene cup, a PET drinks bottle, a polycarbonate lens). Crystallinity, quite literally, is often something you can just look at.
- A tiny ordered speck appears in the cooling melt — a nucleus — either spontaneously or on a stray dust particle or additive.
- Chains begin folding into lamellae that grow radially outward from the nucleus, splaying apart as they lengthen.
- Chain ends, branches, and tangles that will not fit are rejected into the amorphous material filling the gaps between the growing lamellae.
- The ball grows until it bumps into neighbouring spherulites; they impinge and lock together, leaving spherulite boundaries that tile the whole solid.
What dials the crystallinity — and what it buys you
Whether a polymer crystallises 5 percent or 90 percent is decided mostly by two things from earlier guides: how regular the chain is, and how fast you cool it. Regularity comes straight from the chain architecture. A smooth, linear chain with small side groups packs together easily — which is why linear high-density polyethylene (HDPE) reaches perhaps 80 to 90 percent crystallinity. Bolt on branches and they act like burrs that stop chains nesting, so branched low-density polyethylene (LDPE) manages only about 45 to 55 percent. Tacticity matters just as much: an isotactic or syndiotactic chain, with its side groups arranged in a regular pattern, crystallises well, while an atactic chain with side groups thrown on at random (atactic polystyrene, for one) cannot register at all and stays essentially fully amorphous. Crosslinking blocks crystallisation too — the chains are tied down and cannot line up.
Cooling rate is the second dial: cool slowly and chains have time to fold and pack, giving more crystallinity; quench fast and you freeze in more of the tangle. And the payoff is large, because the crystalline regions are denser and stiffer than the amorphous ones. Packing chains regularly squeezes out free volume, so density rises directly with crystallinity — for polyethylene the amorphous phase is about 0.855 g/cm^3 and the fully crystalline phase about 1.00 g/cm^3, and a real sample sits in between like a see-saw balanced between the two (the very rule-of-mixtures logic from the composites and phase-diagram rungs). An LDPE at 0.92 g/cm^3 works out to roughly half crystalline; an HDPE at 0.96 g/cm^3 to about three-quarters. More crystallinity also means more stiffness (Young's modulus can rise several-fold), more strength, more hardness, and more resistance to solvents — bought at the cost of transparency and some toughness.
The glass transition: the amorphous half freezes
Even the crystalline champions still carry amorphous regions, and every fully amorphous polymer is all amorphous — so what happens to that disordered fraction as it cools? It passes through the glass-transition temperature, Tg, exactly the same bend you met for inorganic glass one rung back, now happening to tangled chains. Above Tg the chain segments have enough thermal wiggle to slither past one another: the amorphous material is soft, compliant, rubbery — leather-like. Cool below Tg and that segmental motion freezes out; the tangle locks rigid and glassy, hard and often brittle. Warm caramel that stretches versus cold caramel that snaps is the whole idea, and it is set by crossing one temperature.
This is why two plastics sitting on the same shelf at the same room temperature feel utterly different. A polyethylene bag is limp and flexible because PE's Tg is around minus 110 degrees C — at room temperature its amorphous chains are far above Tg, deep in the rubbery state, and only the crystallites keep the film together. Rigid PVC pipe is stiff and hard because PVC's Tg is around 80 to 85 degrees C — at room temperature it sits below Tg, glassy and rigid. The same idea explains atactic polystyrene (Tg near 100 degrees C): glassy, hard, and brittle at room temperature, which is why a disposable PS cup cracks rather than bends. Move the temperature across Tg and any of these would swap character.
Melting, and the two-temperature map
The crystalline regions have their own transition: the melting temperature, Tm, where the folded lamellae come apart and the ordered chains flood into a disordered liquid. Unlike Tg, this is a true first-order transition with a latent heat of fusion — but because a polymer's crystals vary in thickness and perfection, Tm is smeared over a range of several degrees rather than the razor-sharp point a pure metal shows. A semicrystalline polymer therefore has both temperatures: a Tg for its amorphous fraction and a higher Tm for its crystalline fraction, always with Tm above Tg. As a rough rule of thumb, Tg in kelvin often lands somewhere between about half and two-thirds of Tm in kelvin — handy for a first guess, but with real exceptions, so treat it as a hint, not a law.
STIFFNESS vs TEMPERATURE (log of Young's modulus, as you heat)
stiff |========\ glassy: hard, often brittle
(log E)| \
| \__ drop at Tg (amorphous chains unfreeze)
| Tg \________
| \ rubbery / leathery plateau
| \ (crystallites still hold it)
| \___ drop at Tm (crystals melt)
soft | Tm \_______ liquid: it FLOWS
+-------------------------------------------> temperature
Tg Tm
Semicrystalline (above): TWO drops -- a Tg and a Tm.
Fully amorphous : ONE drop at Tg, then it flows (no Tm).
Crosslinked thermoset : a Tg, then a FLAT rubbery shelf -- never melts.That map ties the three behaviour classes from guide 3 back to just two temperatures. A thermoplastic can be reshaped because heating it past Tm (or past Tg, if it is amorphous) lets the chains flow — which is exactly why PET (Tm about 260 degrees C) can be blow-moulded into bottles and nylon (Tm about 260 degrees C, its high melt point propped up by chain-to-chain hydrogen bonds) spun into fibres. An elastomer is a lightly crosslinked amorphous polymer used above its Tg: natural rubber's Tg near minus 70 degrees C is why it is springy at room temperature — and why a rubber O-ring chilled below its Tg turns glass-hard and shatters, the lesson written into the Challenger disaster. A thermoset like epoxy is a covalent network: it has a Tg (above room temperature, so it is rigid) but no Tm at all — heat it and it chars rather than melts.
How does anyone know a given polymer's Tg and Tm? The workhorse is differential scanning calorimetry (DSC): heat a tiny sample at a steady rate and watch its heat flow. Tg shows up as a small step in the baseline (the heat capacity jumps when segments start moving); Tm shows up as a sharp endothermic dip (heat pouring in to melt the crystals), and the area under that dip even reveals the percent crystallinity. But notice the deeper thread running through this whole guide: everything hinged on whether chains had time to move — to fold, to unfreeze, to flow. That time-and-temperature dependence is not a footnote; it is the defining behaviour of polymers, the reason they act part-solid and part-liquid at once. Pinning it down is viscoelasticity, the whole subject of the final guide in this rung.