A photograph that should not exist
On the morning of 8 April 1982, Dan Shechtman put a rapidly quenched aluminium-manganese alloy into a transmission electron microscope and took a diffraction pattern. What came back stopped him cold: ten bright, razor-sharp spots arranged in a perfect ring, so that rotating the sample by 36 degrees mapped the pattern exactly onto itself. He wrote in his notebook, in English, '10 Fold???'. To feel why those three question marks matter, recall two things you already know. From the diffraction rungs, a pattern of SHARP spots is the signature of long-range order — atoms sitting in a coherent, disciplined arrangement over enormous distances; a disordered glass gives only broad diffuse halos, never crisp spots. And from the previous rung, this is exactly the material that should have been a frustrated glass, its 5-fold-loving icosahedral clusters trapped without any long-range order at all.
So the photograph was a contradiction with teeth. Sharp spots insist the material is superbly ordered over long range. But a ten-fold ring implies a 10-fold (and therefore 5-fold) rotation axis — and the crystallographic restriction theorem proves that no periodic lattice can carry a 5-fold axis. A five-sided pattern simply cannot repeat by translation. You are cornered into a choice: either the material is not really ordered (but the spots are knife-sharp, so it is) or the material is ordered but NOT periodic. That second, once-unthinkable option is the answer, and it has a name: the quasicrystal — a solid with genuine long-range order that never exactly repeats, proudly wearing a symmetry the rules of periodic crystals had declared impossible.
PERIODIC LONG-RANGE DIFFRACTION ALLOWED
(repeats)? ORDER? gives... ROTATIONS
--------------------------------------------------------------------------
CRYSTAL yes yes sharp SPOTS 1,2,3,4,6 only
GLASS / liquid no no broad HALOS none (isotropic)
QUASICRYSTAL NO YES sharp SPOTS 5, 8, 10, 12 too!
--------------------------------------------------------------------------
The shock in one line: sharp spots (= long-range order) with NO periodicity,
arranged in a 10-fold ring that the crystallographic restriction forbids
for anything that repeats. A quasicrystal sits in the empty box no periodic
crystal and no glass could ever fill.Why a repeating pattern cannot be five-sided
It is worth pausing on WHY 5-fold is forbidden, because the quasicrystal is nothing but the escape from that one rule. The crystallographic restriction theorem states that a periodic lattice can carry rotation axes of only five orders: 1-, 2-, 3-, 4-, and 6-fold — never 5-, 7-, 8-fold or higher. This is not a shortage of data or a gap waiting to be filled. Remember the earlier rungs: the 14 Bravais lattices, 32 point groups, and 230 space groups are EXACT, complete enumerations of every way atoms can repeat periodically in three dimensions. A 5-fold periodic crystal is not merely undiscovered — it is provably impossible, in the same way there is no largest prime number. So a symmetry outside that list is a forbidden symmetry, and a sharp 10-fold pattern is a receipt showing you have left the periodic world entirely.
- Picture tiling a floor with identical regular tiles that must meet corner-to-corner and leave no gaps — this is the everyday meaning of a periodic pattern repeating around a point.
- The tiles fit only if their interior angle divides 360 degrees a whole number of times. Squares (90 degrees, 4 around a point), triangles (60 degrees, 6 around), and hexagons (120 degrees, 3 around) all work perfectly.
- Now try a regular pentagon. Its interior angle is 108 degrees, and 360 divided by 108 is 3.33..., not a whole number. Three pentagons cover only 324 degrees and leave a 36-degree wedge open; a fourth overlaps. There is no way to close the ring.
- The same divisibility test fails for 7-, 8-, 9-fold and beyond; only 1, 2, 3, 4, 6 survive. So a translational lattice can never host a 5-fold axis — and to build one, you must give up exact repetition. That surrender is precisely what a quasicrystal makes.
That floor-tiling picture is the whole secret in miniature. The reason your bathroom uses square or hexagonal tiles and never pentagons is the very reason ordinary crystals never show 5-fold rotation axes: a shape that leaves gaps cannot repeat. So the question a quasicrystal answers is beautifully sharp: can you cover the plane completely, with a pattern that is fully ordered and shows 5-fold symmetry, IF you are willing to give up the demand that it repeat? For decades the assumed answer was no. It is yes — and the proof is a tiling made of not one tile but two.
The Penrose tiling: never repeating, never random
In the 1970s the mathematician Roger Penrose found the picture we needed. Take just two tile shapes — a fat rhombus and a thin rhombus, or in the famous version a 'kite' and a 'dart' — and stamp on their edges a set of matching rules that say which edge may sit against which. Obey the rules and the tiles cover the whole plane completely, with no gaps and no overlaps, in a pattern that has clear 5-fold symmetry and yet NEVER repeats: slide the whole Penrose tiling in any direction and it will never land back on itself. This is the two-dimensional portrait of a quasicrystal. The reader climbing this ladder should hold the physical image the whole rung has promised: a quasicrystal tiles space like Penrose kites and darts — never periodic, yet never disordered either.
What keeps it ordered rather than random is a single irrational number: the golden ratio, tau = (1 + square root of 5) / 2 = 1.618..., which is stitched into the fivefold geometry everywhere. Count the tiles in an infinite Penrose tiling and the ratio of kites to darts is exactly tau. That is already the fingerprint of aperiodicity: any pattern that truly repeated would have a whole-number ratio of its two tiles, but tau is irrational, so no repeat is possible. The tiling is also SELF-SIMILAR — group the tiles cleverly and they merge into larger kites and darts forming the same pattern scaled up by tau, a process called inflation you can run forever in either direction. And it is locally homogeneous: any finite patch you point to, however large, recurs infinitely often elsewhere. That is the exact meaning of order without periodicity — deterministic and everywhere-familiar, yet globally never the same twice.
The very simplest cousin lives in one dimension and makes the arithmetic vivid: the Fibonacci chain. Build a line of just two segment lengths, Long and Short, by the rule that each Long splits into Long-Short and each Short becomes a Long. Starting from L you generate L, LS, LSL, LSLLS, LSLLSLSL, ... The counts of L and S follow the Fibonacci numbers 1, 1, 2, 3, 5, 8, 13, 21, and their ratio (2/1, 3/2, 5/3, 8/5, 13/8 = 1.625, 21/13 = 1.615, ...) marches steadily toward tau. The sequence never repeats, yet it is completely determined — and if you shine X-rays down such a chain it produces SHARP diffraction peaks, the one-dimensional echo of Shechtman's spots. Quasiperiodic, ordered, aperiodic: all three at once.
The shadow of a higher dimension
How can a structure be perfectly ordered yet never repeat? The deepest answer is startling and clean: quasiperiodic order is the SHADOW of a periodic lattice living in a higher dimension. Take an ordinary periodic square lattice of dots in 2D, draw a straight line across it at an irrational (golden) slope, and project the nearby dots down onto that line. Because the slope is irrational the line never passes through two dots the same way twice, so the spacings it collects never repeat — and out drops the Fibonacci chain. This 'cut-and-project' trick is the engine of the whole subject: the higher structure is boringly periodic, but its irrational slice is quasiperiodic. A real icosahedral quasicrystal is precisely a 3D slice of a PERIODIC lattice in six dimensions.
This higher-dimensional bookkeeping pays off directly at the diffractometer. Recall that a periodic crystal's diffraction spots ARE its reciprocal lattice, and every spot can be labelled by just three integers (hkl) — three numbers because three basis vectors span the reciprocal lattice. A quasicrystal's spots are still sharp and discrete, but they are dense, filling reciprocal space at every scale, and no three integers can index them all. An icosahedral quasicrystal needs SIX integers per spot (the six reciprocal vectors of that 6D lattice, projected down); a decagonal one, periodic along one axis but quasiperiodic in the plane, needs five. This extra-index scheme, and the higher-dimensional 'superspace' it comes from, is the standard modern language for every aperiodic crystal — the same superspace machinery the next guide will use for incommensurate and modulated structures.
Rewriting the word 'crystal'
A discovery this stubborn forced a definition to change. For a century a crystal MEANT a solid built by periodically repeating a unit cell; quasicrystals broke that definition simply by existing and diffracting sharply. In 1992 the International Union of Crystallography cut the knot by redefining the word from the outside in: a crystal is now any solid whose diffraction pattern is essentially discrete — that is, made of sharp spots. Periodicity was quietly dropped from the definition and demoted to a special case. Periodic crystals become the familiar subset; quasicrystals join them as the aperiodic members. The umbrella term for this widened family is the aperiodic crystal, and it is the true subject of this entire capstone rung.
So what are quasicrystals, as actual matter you could hold? Nearly all are intermetallic alloys. Shechtman's first Al-Mn was metastable, but soon came thermodynamically STABLE quasicrystals grown as beautiful faceted grains — Al-Cu-Fe, Al-Pd-Mn, and even a binary Cd-Yb. They come in two main flavours: icosahedral (quasiperiodic in all three directions) and decagonal (a stack of quasiperiodic planes, periodic along the stacking axis). Their structure gives them strange, useful properties: many are hard, brittle, slippery with famously low friction, and poor conductors of heat and electricity despite being alloys of good metals — a vivid structure-property lesson in how arrangement, not just composition, sets behaviour. And nature got there first: a genuine icosahedral quasicrystal, icosahedrite, was found inside a meteorite, formed without any laboratory.
Now the loop with the previous rung closes cleanly. A liquid whose atoms most want to form 5-fold icosahedral clusters faces a dilemma the crystallographic restriction set for it: those clusters cannot tile a periodic lattice. Two ways out exist. Freeze the conflict UNRESOLVED and you get a metallic glass — icosahedral order trapped locally, no long-range order at all. Resolve it by keeping the 5-fold order but SURRENDERING periodicity, and you get a quasicrystal — long-range order, forbidden symmetry, no repeat. That single choice is the doorway into the rest of this capstone: structures that stretch or break periodicity while staying rigorously ordered — the incommensurate and modulated crystals next door, the giant-celled complex intermetallics and high-entropy alloys, and the framework materials beyond, all of them now living comfortably under the widened word 'crystal'.