Two ways to be a solid
The last rung ended with atoms clinging together in a bonding well. Now we ask the very next question: once they stick, how do they line up? Nature offers only two answers. In a crystalline solid the atoms sit in a pattern that repeats, unbroken, in every direction — like soldiers on a parade ground, so regular that if you stood on any one atom and looked around, then walked to an identical atom far away and looked again, the view would be exactly the same. This is called long-range order: the arrangement is predictable not just next door but a million atoms away. Nearly all metals, most ceramics, and many minerals are crystalline.
The other answer is disorder. In an amorphous solid — window glass is the classic — the atoms are still bonded, still packed close, but with no repeating pattern: freeze a swirling crowd exactly where it stands and you have the picture. The everyday word for such a material is a glass, and the everyday way to make one is to cool a liquid so fast the atoms never get a chance to file into rank before they seize up. A crystal is a liquid that took its time and organised; a glass is a liquid that was ambushed by the cold. Same atoms, same bonds — the only difference is whether order reaches across the whole solid or dies out after a few neighbours.
The repeating block: lattice and unit cell
If a crystal repeats forever, we do not need to describe every atom — we only need to describe the pattern and the stamp that copies it. Strip a crystal down to bare geometry, replacing each repeating atom (or group of atoms) with a single point, and you get a space lattice: an infinite, perfectly regular grid of points in three dimensions. It is the wallpaper pattern of the crystal with the ink removed, just the ghostly framework where atoms will hang.
You never draw the whole infinite grid. Instead you find the smallest chunk that, stacked side by side like identical bricks with no gaps and no overlaps, rebuilds the entire crystal — the unit cell. It is the crystal's repeat unit, the tile of the wallpaper. Once you know the unit cell, you know the crystal: the rest is just copy-and-paste. It turns out that only seven distinct cell shapes are needed to catalogue every crystal — the seven crystal systems, sorted by their edge lengths and corner angles (cubic is the tidiest, with three equal edges meeting at right angles). Allow the lattice points to sit not only at the cell corners but also at face centres or the body centre, and the seven shapes give rise to exactly fourteen distinct patterns, the fourteen Bravais lattices. Fourteen — that is the complete list of ways to fill space with a repeating point grid. Guide 2 lives inside this cell.
The three structures a metal picks
Because a metal's bond has no direction — recall the electron sea from the last rung — its atoms behave like hard balls that just want to pack together as tightly and simply as possible. Three arrangements win almost every time. Face-centered cubic (FCC) puts an atom at each cube corner and one in the middle of each face; body-centered cubic (BCC) puts one at each corner and a single atom dead in the centre; and hexagonal close-packed (HCP) stacks hexagonal layers in an ABAB rhythm. Copper, aluminium, and gold are FCC; iron at room temperature and tungsten are BCC; zinc, magnesium, and titanium are HCP.
structure atoms/cell coordination # packing factor examples --------- ---------- -------------- -------------- ----------------- FCC 4 12 0.74 Cu, Al, Au, Ni BCC 2 8 0.68 Fe(RT), W, Cr HCP 6* 12 0.74 Zn, Mg, Ti, Co coordination # = how many nearest neighbours touch each atom packing factor = fraction of the cell actually filled by atom-spheres FCC & HCP both hit 0.74 -- the tightest that equal spheres can ever pack (* HCP conventional cell contains 6 atoms; its primitive cell has 2)
Two numbers in that table earn their keep. The coordination number counts how many neighbours touch each atom — 12 for the close-packed FCC and HCP, only 8 for the looser BCC — and the atomic packing factor is the fraction of the cell that atom-spheres actually fill. FCC's 0.74 is not a coincidence you have to memorise; it is provably the densest that identical spheres can ever be packed, the same 74 percent a greengrocer reaches stacking oranges. These aren't idle facts: how tightly atoms pack and how many neighbours each has will later govern how easily the metal deforms, how dense it is, and how it transforms when heated. We compute all three — coordination, packing, and the theoretical density that falls out of them — in guide 3.
One element, more than one crystal
Here is a fact that quietly runs half of metallurgy: the same atoms can crystallise into more than one structure, and which one wins depends on temperature and pressure. This is polymorphism (called allotropy when we're talking about a pure element). Iron is the star example. Cool and calm at room temperature it is BCC (metallurgists call this form ferrite); heat it past about 912 degrees C and the very same iron atoms rearrange into FCC (austenite); push higher still and it flips back to BCC. That single, reversible shuffle between BCC and FCC is the loophole that makes steel hardenable — nearly everything in the later heat-treatment rungs, from quenching to the iron-carbon diagram, exists because iron cannot make up its mind about its crystal structure.
Carbon makes the point even more dramatically. The same carbon atoms build both diamond — where each carbon covalently bonds to four neighbours in a rigid three-dimensional cage, giving the hardest natural material — and graphite, where carbon bonds strongly within flat sheets but the sheets are held to each other only by weak secondary bonds, so they slide apart and smear onto paper as pencil 'lead'. Diamond and graphite are the same element at the same price of atoms, yet one is the hardest thing you own and the other is a lubricant. If you ever doubted that structure, not just composition, is what decides a material's fate, let diamond and graphite settle it: rearrange the very same atoms and you get a gemstone or a pencil.
Naming directions, single crystals, and grains
Once a lattice repeats in space, we need a shared language to point at a particular direction or a particular plane of atoms inside it — otherwise 'that diagonal slice' is hopeless to communicate. That language is Miller indices: a compact set of small whole numbers, written like (111) for a plane or [110] for a direction, that names any plane or line in the crystal unambiguously. It matters because atoms are not spread evenly in every direction — some planes are densely packed with atoms, others sparse — and a metal deforms most easily along its most crowded, most closely packed planes. Guide 4 is devoted to reading and writing these indices; for now just know that crystals have a built-in coordinate grammar, and it has real physical consequences.
So far we have imagined a crystal whose pattern runs unbroken from edge to edge — a single crystal, and they do exist (a silicon wafer, a turbine blade, a gemstone). But melt most metals and let them freeze and you get a polycrystal instead: crystallisation starts in many places at once, each little seed growing its own crystal in its own random orientation, until the growing crystals collide. Each patch is a grain, and where two mismatched grains meet is a grain boundary — a thin, disordered seam of atoms that fit neither pattern cleanly, exactly like floor tiles laid in two rooms at different angles, jammed against each other along a ragged join. A typical piece of metal is a mosaic of thousands or millions of such grains.
How we know: X-rays and Bragg's law
A fair objection: atoms are far too small to see, so how does anyone know they sit in these tidy patterns at all? The answer is that we don't look — we listen for echoes. Shine X-rays (whose wavelength happens to be about the spacing between atomic planes, roughly 0.1 nm) onto a crystal, and the regularly spaced planes of atoms scatter the waves. In most directions the scattered wavelets cancel, but at a few sharp angles they line up crest-on-crest and reinforce, throwing back a bright beam. Those special angles obey Bragg's law, n(lambda) = 2 d sin(theta), which links the X-ray wavelength (lambda) and the reflection angle (theta) directly to d, the distance between the atomic planes. Measure the angles where the beams flash out, and you can solve backwards for the spacing of planes you could never see.
This is the fingerprint reader of materials science. A crystalline solid gives a pattern of sharp bright spots or peaks — each one a plane spacing announcing itself — and the whole pattern is a signature you can match to identify the phase, measure the unit cell, even tell FCC iron from BCC iron. An amorphous solid, having no repeating planes, gives no sharp peaks at all, only broad, blurry humps — the diffraction pattern itself confirms the difference between order and disorder we opened with. Everything the next four guides claim about unit cells, packing, and planes was learned this way, by reading the echoes X-rays throw back. That closes the loop: we started by asserting atoms line up in patterns, and Bragg's law is how we earned the right to say so.