atom-probe tomography
Atom-probe tomography (APT) comes closer than any other technique to a science-fiction dream: it takes a real piece of material apart one atom at a time, records where each atom came from in three dimensions, AND identifies which chemical element each one was. The other microscopes in this field give you either a projected image or a surface; APT gives you a genuine three-dimensional map of a small volume with both the positions and the identities of the atoms filled in. It is the way to answer questions like 'is there a single layer of one element segregated to this grain boundary?' or 'how are the dopant atoms clustered inside this alloy?'
The method works by controlled, atom-by-atom demolition. You first shape the specimen into an extremely sharp needle, its tip only tens of nanometres across (usually carved with a focused ion beam). Cooled to cryogenic temperature in ultrahigh vacuum, the needle is held at a high voltage; the electric field concentrates so intensely at the sharp apex that it is on the verge of ripping atoms off. A short pulse — of extra voltage or of laser heat — then tips the balance and makes the outermost atom leave as an ion, one at a time. Each departing ion flies to a position-sensitive detector: where it lands records its original (x, y) position on the tip, and how long it took to arrive (its time of flight) reveals its mass-to-charge ratio and hence its element. Strip the needle away layer by layer and a computer stacks the results into a full three-dimensional, chemically-labelled reconstruction, atom by atom.
No other method combines near-atomic spatial resolution with element identification in three dimensions, which makes APT invaluable for nanoscale chemistry: dopant distributions in semiconductors, solute clustering in steels, and the composition of interfaces and precipitates. But be candid about the caveats. It is destructive — you evaporate the very sample you are studying. The detector catches only a fraction of the atoms (efficiency is typically around a half to four-fifths), so the map is a representative sample, not every atom. The reconstruction assumes a smooth tip shape and so carries geometric distortions, with spatial accuracy noticeably better along the depth direction than across it. And preparing that flawless needle is demanding, especially for brittle or multiphase materials.
Investigating why a steel embrittles, an APT reconstruction of a needle spanning a grain boundary shows a thin sheet of phosphorus atoms, only about one atom thick, decorating the boundary plane. No projected image could prove that a single atomic layer of one element sits exactly there, in three dimensions and identified by species.
APT is the only probe giving 3D atom positions with chemical identity — ideal for boundary and cluster chemistry.
APT is destructive (the sample is field-evaporated away) and detects only a fraction of atoms — efficiency is roughly a half to four-fifths — so the map is representative, not complete. The reconstruction also has geometric distortion, with better accuracy in depth than laterally.