Why light hits a wall — and electrons blow past it
Guide 1 handed us grains under an optical microscope: section, mount, grind, polish, etch, and the metallography lays a whole microstructure open to visible light. But keep turning the magnification knob and the picture does not keep improving — past a point it simply blurs, and no better lens or brighter lamp helps. That ceiling is not a flaw in the glass; it is diffraction. Any wave can only resolve detail down to roughly its own wavelength, and visible light is 400 to 700 nm across, so two features closer than about 0.2 micrometres (200 nm) smear into one. Grains, coarse pearlite, and big pores live comfortably above that wall. Individual dislocations, nanometre precipitates, and the fine lamellae of hardened steel live below it, invisible. To go deeper we need a far shorter wave.
Here is the beautiful trick: electrons are waves too. An electron has a de Broglie wavelength lambda = h/p, and the faster you fire it the shorter that wavelength gets. Accelerate an electron through a voltage V and its wavelength is roughly 1.23/sqrt(V) nanometres, with V in volts. Feed in the numbers: at a modest 20 kV (20000 V) you get 1.23/141 which is about 0.0087 nm — already smaller than a single atom. At 200 kV, once you correct for relativity, the wavelength shrinks to about 0.0025 nm. That is 10,000 to 100,000 times shorter than a photon of light. An electron beam is a wave fine enough, in principle, to see atoms.
The SEM: painting a surface with a beam
The scanning electron microscope does not project a picture the way a light microscope does. It focuses the beam to a fine spot and sweeps it across the specimen line by line, exactly the way an old CRT television painted its screen. At every point it pauses, measures how many electrons come back off the surface, and sets one pixel to that brightness. The image is therefore a map of a measured signal, built dot by dot, not a shadow cast through a lens. Magnification is just the ratio of the scanned area to the display — shrink the sweep and you zoom in smoothly. Its signature virtue is an enormous depth of field: because the beam is a thin pencil, steep peaks and deep valleys stay in focus together, which is why SEM images of pollen, insect eyes, and fracture surfaces look so vividly three-dimensional.
Two kinds of electron come back, and they tell different stories. Secondary electrons are low-energy stragglers knocked loose from the sample; only those born in the top few nanometres can crawl out, so they carry topography — edges and tips glow brighter (the 'edge effect'), giving that crisp surface-relief image. Backscattered electrons are beam electrons that ricochet straight back off atomic nuclei; a heavier atom scatters more of them, so brightness climbs with atomic number and heavy phases shine pale against a light-element background. That is composition contrast for free. Depth of field plus surface detail is why the SEM owns fractography: on a broken part it reads the ductile dimples, the flat cleavage facets and river lines of a brittle fracture, and the one-per-cycle striations of fatigue we met in the failure rung — the fingerprints of how the thing died.
focused electron beam (spot ~1 nm wide)
|||
===================|||=================== SPECIMEN SURFACE
Auger e- \|/ <- top ~1 nm
secondary e- V <- ~5-50 nm => TOPOGRAPHY (sharp)
.-'|'-.
/ BSE \ <- ~0.1-1 um => Z / COMPOSITION
/ region \
| X-rays | <- ~0.5-3 um => EDS: which elements
\ (EDS) / (fuzzy: whole pear glows, not spot)
\ /
'-.__.-'
the pear-shaped INTERACTION VOLUME
beam spot is nm-sharp, but each signal escapes from a different depth:
SE = shallow (surface shape) | BSE = mid (composition) | X-ray = deep (EDS)The SEM also answers 'what is it made of', through energy-dispersive X-ray spectroscopy (EDS). When the beam knocks an electron out of an atom's inner shell, an outer electron drops in and emits an X-ray whose energy is a fixed fingerprint of that element. An EDS detector sorts the X-rays by energy and reads off which elements are present and roughly how much. So SEM plus EDS gives you shape and composition together — point at a suspicious particle and ask what it is. Be honest about the limits, though: light elements (hydrogen, helium, lithium, often boron and carbon too) are hard or impossible to detect, the numbers are only semi-quantitative, and the chemistry is blurrier than the picture because the X-rays come from that whole micron-wide pear, not the spot. And the sample must survive vacuum and conduct electrons, so ceramics, polymers, and biology usually get a whisper-thin sputter-coat of gold or carbon first.
The TEM: shining electrons through a slice
The transmission electron microscope flips the geometry. Instead of bouncing electrons off a surface, it fires them clean through the specimen, the way a projector shines light through a slide, and forms the image from what makes it out the other side. The price of that is brutal: the sample must be electron-transparent, thinned to under about 100 nm and often under 50 — a hundredth the thickness of a soap film. Getting there is the real work of TEM, done by electropolishing, ion milling, or lifting out a sliver with a focused ion beam, and it is slow, skilled, and destructive. So be clear-eyed: a TEM shows you a tiny, painstakingly prepared foil, not a bulk surface, and whether that sliver represents the whole part is a question you must always ask.
Why endure all that? Because of what appears. As electrons cross the thin crystal, the planes of atoms diffract them, and anything that bends those planes changes the local diffraction — so lattice defects paint themselves onto the image as dark lines and blobs. This is where the dislocation, that invisible line defect we leaned on all through the strengthening rung to explain slip and work hardening, finally becomes something you can literally photograph: tangled forests of dislocations, pile-ups jamming against a grain boundary, the ruck in the rug caught in the act. You can count their line-length per unit volume — the dislocation density that climbs as a metal is cold-worked — and watch precipitates, stacking faults, and twins that are far too small for any light microscope. Turn a knob and the TEM instead shows the diffraction pattern itself: sharp spots for a single crystal, rings for many grains — an atomic ruler we will pick up properly in guide 3.
Fingertips instead of light: AFM and STM
There is a third family that throws out beams and lenses entirely. A scanning-probe microscope drags an atomically sharp needle back and forth across the surface and builds a map from what the tip feels at each point — less like taking a photograph and more like a blind person reading Braille with one fingertip, or a record-player stylus tracing a groove. There is no wavelength limit at all here, because nothing is imaging with a wave; the resolution is set by how sharp you can make the tip and how gently you can move it.
The two members split by what the tip senses. The scanning tunnelling microscope (STM) measures the quantum tunnelling current that leaks across the tiny gap between tip and surface; that current changes so steeply with distance that it resolves single atoms — but only on a clean, conducting sample. The atomic force microscope (AFM) instead feels the faint force between tip and surface through the bending of a soft cantilever, read by a laser bounced off its back, so it works on insulators too, in air or even under liquid. What these add that electron microscopes cannot is a genuinely calibrated three-dimensional height map: an SEM gives you a picture that looks 3D but has no true vertical scale, whereas an AFM measures depth to a fraction of a nanometre. Their honest limits are a small scan area (microns at most), slow going, and tip-shape artifacts — a blunt tip smears every feature into its own fat outline, and neither can see below the surface.
What each tool resolves — and is for
None of these instruments is simply 'better' — each answers a different question at a different length scale, and the skill is matching the tool to what you actually need to know. Light microscopy (guide 1) reads grains and phases at the micron scale, cheaply and over large areas, and is where nearly every investigation starts. The SEM covers roughly 1 nm to millimetres: surfaces, fracture faces, and, with EDS, composition maps. The TEM reaches inside a thin foil to dislocations, precipitates, and single atomic columns. Scanning probes give calibrated topography down to the atom, on the bench, in air. Ask first how big your question is, then reach for the instrument whose window straddles that size.
These are the imaging tools, and they are complementary, not rivals — a real failure investigation walks steadily down the ladder, eye to hand-lens to light microscope to SEM to TEM, zooming in only where the last picture left a question. But an image, however sharp, mostly tells you shape; it does not always tell you which crystal phase you are looking at or which molecules are bonded to what. That is the next rung of tools. The TEM's diffraction pattern was a taste of guide 3, where X-ray diffraction identifies phases, measures lattice spacings, and reads residual stress across a whole bulk sample without any thinning, and where the spectroscopies — the EDS you just met, plus surface XPS and Auger, and molecular Raman and infrared — pin down chemistry and bonding. Guides 4 and 5 then add thermal analysis and nondestructive testing. Seeing is only the first half of characterization; putting the right number on what you see is the rest.