The SEM Field Guide

Notebook · Troubleshooting

Astigmatism, or: why your image smears one way, then the other

The defect every operator meets in their first week and fights, on and off, for the rest of their career.

Focus through a feature on a poorly stigmated column and you'll see the signature: just below focus, everything smears along one direction, as if the image were dragged. Pass through best focus — briefly, deceptively sharp-ish — and the smear returns, rotated ninety degrees. Round particles render as crosses or stars. Fine texture looks combed. And no amount of focus knob will fix it, because focus is not the problem.

The problem is that the beam's cross-section isn't round. An ideal column focuses the beam to a symmetric point; a real column has slightly elliptical lens fields — from manufacturing tolerances, from contamination on apertures, from the sample's own charging — so rays in one plane focus at a different height than rays in the perpendicular plane. Between those two focal planes the "point" is actually an ellipse whose long axis flips 90° as you pass through. Your image is sharp in one direction and soft in the other, and which direction depends on which side of focus you're on.

Astigmatic beam underfocus — smear ↔ "best" focus — disc of least confusion overfocus — smear ↕ Corrected beam round on both sides of a tight focus — blur, but symmetric blur
Fig. 1 — The tell. An astigmatic beam is an ellipse that rotates 90° through focus; a corrected beam is round everywhere, merely larger or smaller. Directional smear that flips with the focus knob is astigmatism, every time.

The stigmators, and a routine that converges

Every SEM carries a pair of correctors — stigmator X and Y — that impose a compensating ellipticity of adjustable strength and orientation. The correction itself is easy. What's hard, at first, is that focus and stigmation interact: each stigmator tweak shifts best focus slightly, so operators chase their own tail. This routine converges:

  1. Find a feature with detail in all directions — a rough particle, an edge with texture. Avoid long straight edges; they hide smear parallel to themselves.
  2. Focus as well as the image allows at your working magnification.
  3. Go up in magnification — 2× to 5× beyond where you'll actually image. Corrections invisible at 5,000× are glaring at 20,000×.
  4. Adjust stigmator X for maximum sharpness. Not "less smear" — maximum crispness. Then stigmator Y. Then refocus.
  5. Repeat the X–Y–focus cycle once or twice. Each pass should need less. If the image "pops" into a new tier of sharpness, you're done; return to working magnification, where the image should now look effortless.

When it won't stay fixed

Stigmation that drifts session to session, or won't null out at reasonable corrector strength, is a message. The usual senders, in order of likelihood: a contaminated aperture (debris charges and acts as a weak, wandering lens — clean or replace it), a charging sample pulling the beam around (an astigmatism-shaped symptom with a sample-prep solution), large changes in working distance or kV since the last correction (routine; just re-stigmate), and, rarely, genuinely misaligned column elements. Exhaust the cheap explanations before the expensive ones: the aperture strip has taken the blame for many a service call it deserved.

Astigmatism never stops being true. It waits for a new aperture, a change of working distance, a humid week, a charging sample. The operators who seem never to fight it are simply the ones who correct it so habitually — thirty seconds, every session, at the start — that they've forgotten it's a fight.