From reciprocal space back to real space
Look back over this rung. Every method so far read structure through scattered waves — selected-area and convergent-beam patterns and Kikuchi lines used electrons, HRTEM and STEM reconstructed atomic columns from an electron wave passing through a thin foil, EBSD mapped orientation from backscattered electrons, and neutron diffraction reached the light atoms and magnetic order that X-rays miss. All of them are, at heart, forms of structure characterization by diffraction: you send in a wave, you read the pattern it leaves, and you infer the atoms. This final guide meets the atoms in a completely different way — in real space, up close, by touching them.
This is the great divide. A diffraction experiment works in reciprocal space and averages over an astronomical number of unit cells at once, so it hands you the ideal periodic structure with superb statistics — but it cannot tell you where any single atom or defect actually sits. The probes in this guide do the opposite: they work directly in real space over a tiny region, trading long-range averaging for the power to point at one atom, one surface step, one interface. As a bonus they sidestep the phase problem entirely — you see positions directly, so there is no lost phase to guess back — but only across a field of view a few nanometres wide.
Meet the family. The scanning tunneling microscope feels the electron cloud at a surface through a quantum tunnel; the atomic force microscope feels the force between a tip and a surface, so it works even on insulators; and atom-probe tomography takes a needle-shaped specimen apart one ion at a time, weighing each so it can map atoms in 3D with their chemical identity. Two are surface probes, one dismantles a tiny 3D volume. What unites all three is that they touch rather than photograph — the atom is met, not deduced from a shadow.
The STM: feeling atoms with a quantum tunnel
Bring an atomically sharp metal tip within about 0.5 to 1 nm of a conducting surface and apply a small voltage. Classically no current should flow — there is a vacuum gap in the way. But quantum mechanically, electrons leak across that classically forbidden gap: a tiny tunneling current, typically a fraction of a nanoampere, flows between tip and sample. This is the whole engine of the scanning tunneling microscope. Think of hearing a whisper through a thick wall — inaudible from across the room, but it swells dramatically the instant you press your ear against the plaster.
The magic is how ferociously that current depends on the gap. It decays as exp(-2 times kappa times d), where kappa is about 1 per angstrom for a typical work function near 4 eV. So pulling the tip just 1 angstrom closer multiplies the current by about e^2 — close to a full order of magnitude. That savage sensitivity is exactly why the STM resolves single atoms: the current is dominated by the single frontmost atom of the tip tunneling to the single nearest atom of the surface. Scan the tip line by line, hold the current constant with a feedback loop that adjusts the height, and the height map traces the atomic corrugation of the surface — a real-space picture at the atomic scale.
The price of the tunnel is that the STM demands a conductor (or a thin film on one), usually ultra-high vacuum, and often a cold, atomically clean surface — and it only ever sees that surface, never the interior. But the same exquisite control of a single tip has a spectacular flip side: you can nudge individual atoms across the surface, one at a time, as the famous xenon-atom logos first proved. So the STM is not only a microscope but an atom-scale hand — it images and moves single atoms, something no diffraction method can dream of doing.
The AFM: a nanoscale fingertip on the surface
The atomic force microscope answers the STM's biggest limitation — the need for a conductor. Instead of a tunneling current it measures force. A tiny sharp tip sits on the end of a flexible cantilever; as the tip approaches the surface, interatomic forces (a gentle van der Waals attraction at first, then hard repulsion once electron clouds start to overlap) bend the cantilever, and a laser beam bouncing off its back reads the deflection to a fraction of an angstrom. It is, quite literally, a blind reader's fingertip running across a page of Braille — feeling the bumps rather than looking at them. This is the atomic force microscope.
Because it feels force rather than current, the AFM works on insulators, polymers, ceramics, and living biological samples — and even in air or in liquid, a huge widening from the STM's ultra-high-vacuum-and-conductor world. It runs in several modes: contact mode drags the tip along the surface and feels the repulsion; tapping and non-contact modes oscillate the cantilever and sense the force gradient, gentle enough not to plough through a soft sample. That versatility has made the AFM the everyday nanoscale topographer across materials science and biology alike.
Stay honest, though, about the phrase 'atomic resolution.' A contact-mode 'atomic' image usually shows the lattice periodicity — an averaged corrugation — rather than every individual atom, because many tip atoms touch the surface at once and blur the single-atom detail. Resolving a genuine single vacancy or a lone dopant needs non-contact AFM with a very stiff sensor in ultra-high vacuum, where the force gradient from just the frontmost tip atom dominates. So 'atomic-resolution AFM' and 'seeing every atom, including the missing ones' are not the same claim — a distinction worth keeping straight.
Atom-probe tomography: a crystal taken apart atom by atom
The most radical probe of all does not scan a surface — it dismantles the specimen. Shape the material into an extremely sharp needle, apex radius under about 50 to 100 nm, carved with a focused ion beam. Sit it in ultra-high vacuum, cool it to cryogenic temperature, and raise it to a few kilovolts. Because the field concentrates at a sharp tip, the field at the apex reaches tens of volts per nanometre — of order 10^10 V/m — strong enough to rip the outermost atoms off one at a time. Each surface atom ionizes and field-evaporates, flying off toward a detector. This is atom-probe tomography.
- A short voltage or laser pulse tips the field just over threshold, so exactly one (or a very few) surface atom field-evaporates as a positive ion.
- A position-sensitive detector records WHERE each ion lands; back-projecting its flight path fixes that atom's (x, y) position on the needle.
- The flight time from pulse to detector gives the mass-to-charge ratio — which pins down the chemical element, and often the very isotope, of that single ion.
- Atoms leave layer by layer as the needle sharpens away, so the ORDER of evaporation encodes the depth z of each atom.
- Repeat for tens of millions of atoms; software reconstructs a 3D point cloud in which every point is one atom, coloured by its element.
The payoff is unique among everything in this rung: a three-dimensional map at near-atomic resolution that also carries chemical identity — the one method that gives you composition in space at this scale. That is why atom-probe tomography is unmatched for questions the averaged methods cannot touch: how much boron or phosphorus has segregated into the plane of a grain boundary, how a substitutional impurity or dopant is distributed and clustered inside a semiconductor, or the earliest nanometre-scale precipitates long before they are big enough to throw a diffraction peak. It reads the chemistry of a defect, atom by atom, in 3D.
Choosing the right probe for the structural question
Now step back across this whole rung and the diffraction rung before it, and the choice of instrument organizes along three simple axes. Is your question about the reciprocal-space average or a real-space local spot? About the surface or the bulk? About where atoms sit (structure) or which atoms they are (composition)? No single instrument wins on all three axes at once — you pick the one whose strengths line up with what you actually need to know.
CHOOSING A STRUCTURAL PROBE (space | depth | what it tells you)
THE QUESTION BEST TOOL WHY
------------------------------------ -------------------- ---------------------------
Average crystal structure of a bulk X-ray powder reciprocal-space, huge
solid; identify an unknown phase diffraction averaging, great statistics
Light atoms (H), isotopes, or the neutron neutrons scatter off nuclei
magnetic ordering of the spins diffraction + carry a magnetic moment
Local structure of ONE tiny region, HRTEM / STEM real-space, atomic columns
a defect, atomic columns (+ SAED / CBED) in a thin foil; local
diffraction pattern
Grain orientation + texture mapped EBSD in the SEM backscatter Kikuchi pattern
over a whole mm-scale surface indexed at every pixel
Surface atomic arrangement, its STM / AFM real-space, feels the top
electronics, moving single atoms layer only (surface probe)
3D COMPOSITION near atomic scale: atom-probe field-evaporate a needle,
dopants, clustering, segregation tomography weigh every ion in 3D
RULE OF THUMB: averaged + bulk -> diffraction ; local structure -> TEM ;
surface -> STM/AFM ; chemistry-in-3D -> atom probe.The real lesson of this rung is that these probes are complementary, not rivals — the strongest studies braid them together. A neutron or X-ray pattern gives the ideal averaged unit cell; the TEM zooms in on a defect and its local diffraction; EBSD maps the orientation landscape across a whole surface; the STM and AFM read the true surface termination; and atom-probe tomography delivers the buried 3D chemistry. Each answers precisely what the others are blind to, and a full account of a material's structure comes from stitching the reciprocal-space and real-space views into one picture. This closes the Electron and Neutron Methods rung — and the next rung climbs to microstructure and texture, the aggregate of grains you can now measure with exactly the tools assembled here.