medium-range order
Imagine describing a crowd. Look at any one person and their immediate handful of neighbours — how far apart they stand, which way they face — and you have described SHORT-range order. Now zoom out and see how those little clusters link up: do three or four groups form a ring? Does a chain of clusters snake across the plaza? That in-between scale — bigger than a single cluster but far smaller than the whole crowd — is medium-range order. In an amorphous solid it is the organisation that survives just beyond the nearest-neighbour shell: how the well-defined local units connect to one another over the next one or two nanometres.
Concretely, in silica glass the short-range order is each silicon bonded to four oxygens in a tetrahedron — that part is almost perfectly regular. Medium-range order is the next question: how do those tetrahedra join up? They share corners, but the silicon-oxygen-silicon angle at the shared corner can vary (peaking around 144 degrees), and the tetrahedra can twist relative to each other. Out of this comes a preference for certain ring sizes — chains of tetrahedra that close back on themselves after, say, five or six members — and these rings, spanning perhaps 0.5 to 2 nanometres, are the fingerprint of medium-range order. It shows up in scattering as a faint low-angle feature called the first sharp diffraction peak, a peak whose position corresponds to a repeat distance of about 4 to 5 angstrom even though the material has no true long-range repeat.
Medium-range order matters because it is the layer of structure that most distinguishes one glass from another and controls many properties, yet it is the hardest to pin down. Short-range order is easy to measure (the first peak of the pair distribution function); long-range order is simply absent. Medium-range order lives in the awkward middle — real but subtle, showing up as weak, broad features that are notoriously difficult to interpret. Being honest, our picture of it is still partly model-dependent: researchers debate ring statistics, void distributions, and cluster connectivity precisely because medium-range order is genuinely hard to see.
In amorphous silica the Si-O bond length (short-range order) is fixed at about 1.6 angstrom and hardly varies. But the ring statistics — whether tetrahedra link into five-membered, six-membered, or seven-membered rings — vary from glass to glass and shift when the glass is compressed or the melt is cooled at different rates. That variable ring population is medium-range order, and it is why two silica glasses with identical short-range order can still differ in density and behaviour.
Short-range order sets the local unit; medium-range order sets how those units link over the next nanometre or two.
Do not treat medium-range order as an established, fully measured quantity like a bond length. It is real but hard to observe directly, and much of what we say about it comes from models fitted to weak scattering features — it remains an active, sometimes contested, research topic.