Scanning Electron Microscopy

sem Electron Microscopy Electron Beam Particle
View Benchmarks (1)

SEM forms images by rastering a focused electron beam (1-30 keV) across the specimen surface and collecting secondary electrons (SE, topographic contrast) or backscattered electrons (BSE, compositional Z-contrast). Resolution is determined by the probe diameter (1-10 nm), accelerating voltage, and interaction volume. Key artifacts include charging in non-conductive specimens, drift, and contamination.

Forward Model

Raster Scan Detection

Noise Model

Poisson

Default Solver

direct imaging

Sensor

ELECTRON_DETECTOR

Forward-Model Signal Chain

Each primitive represents a physical operation in the measurement process. Arrows show signal flow left to right.

P e⁻ Electron Beam C probe Probe Scanning D g, η₁ SE / BSE Detector
Spec Notation

P(e⁻ beam) → C(probe) → D(g, η₁)

Benchmark Variants & Leaderboards

SEM

Scanning Electron Microscopy

Full Benchmark Page →
Contribute Dataset Compete
Spec Notation

P(e⁻ beam) → C(probe) → D(g, η₁)

Standard Leaderboard (Top 10)

# Method Score PSNR (dB) SSIM Trust Source
🥇 SwinIR 0.772 33.4 0.930 ✓ Certified Liang et al., ICCVW 2021
🥈 Noise2Void 0.724 31.6 0.895 ✓ Certified Krull et al., CVPR 2019
🥉 BM3D 0.635 28.5 0.820 ✓ Certified Dabov et al., IEEE TIP 2007
4 Wiener Filter 0.503 24.8 0.680 ✓ Certified Analytical baseline
Mismatch Parameters (3) click to expand
Name Symbol Description Nominal Perturbed
beam_energy ΔE Beam energy error (keV) 0 0.1
stigmatism ΔA_s Astigmatism (nm) 0 5.0
working_distance ΔWD Working distance error (mm) 0 0.1

Reconstruction Triad Diagnostics

The three diagnostic gates (G1, G2, G3) characterize how reconstruction quality degrades under different error sources. Each bar shows the relative attribution.

G1 — Forward Model Accuracy How well does the mathematical model match reality?

Model: raster scan detection — Mismatch modes: charging, drift, contamination, astigmatism, vibration

G2 — Noise Characterization Is the noise model correctly specified?

Noise: poisson — Typical SNR: 15.0–40.0 dB

G3 — Calibration Quality Are instrument parameters accurately measured?

Requires: beam current, accelerating voltage, working distance, stigmation, aperture alignment

Modality Deep Dive

Principle

Scanning Electron Microscopy rasters a focused electron beam (0.1-30 keV) across the sample surface. Secondary electrons (SE) emitted from the top few nanometers provide topographic contrast, while backscattered electrons (BSE) from deeper interactions reveal compositional contrast (higher Z → more BSE). The image is formed point-by-point, with resolution down to 1-5 nm determined by the probe size.

How to Build the System

Operate a field-emission SEM (FEG-SEM, e.g., Zeiss GeminiSEM, JEOL JSM-7800F) under high vacuum (< 10⁻⁴ Pa). Mount samples on conductive stubs with carbon tape or silver paint. Non-conductive samples must be sputter-coated (5-10 nm Au/Pd or C) to prevent charging. Set accelerating voltage (1-5 kV for surface detail, 10-20 kV for BSE compositional contrast). Select appropriate detectors (Everhart-Thornley for SE, solid-state for BSE). Align the column and perform astigmatism correction.

Common Reconstruction Algorithms

  • Noise reduction by frame averaging or Kalman filtering
  • Charging artifact compensation (dynamic focus, low-kV imaging)
  • 3-D surface reconstruction from stereo-pair SEM images
  • Deep-learning SEM denoising (for low-dose or fast-scan images)
  • Automated particle analysis and morphometry

Common Mistakes

  • Sample charging causing bright streaks and image distortion
  • Astigmatism not corrected, producing elongated features
  • Excessive beam current damaging or contaminating delicate samples
  • Carbon contamination from residual hydrocarbons in the chamber
  • Wrong working distance causing suboptimal resolution or depth of field

How to Avoid Mistakes

  • Coat non-conductive samples or use low-vacuum/variable-pressure mode
  • Correct astigmatism carefully using the wobbler on a recognizable feature
  • Use the minimum beam current needed; work at low kV for beam-sensitive samples
  • Plasma-clean the chamber and samples; use a cold trap to reduce contamination
  • Optimize working distance for the specific detector and resolution requirement

Forward-Model Mismatch Cases

  • The widefield fallback applies optical Gaussian blur, but SEM image formation involves electron-sample interaction (secondary electron yield depends on surface topography and composition) — the contrast mechanism is fundamentally different from optical fluorescence
  • SEM contrast (SE and BSE signals) depends on accelerating voltage, material Z-number, surface tilt, and detector geometry — the widefield PSF convolution model cannot capture these electron-matter interaction physics

How to Correct the Mismatch

  • Use the SEM operator that models the electron probe profile (sub-nm spot) and secondary/backscattered electron yield as a function of local surface topography and composition
  • Include the interaction volume (Monte Carlo electron trajectory simulation), detector angular acceptance, and signal mixing between SE (topography) and BSE (composition) channels

Experimental Setup

Instrument

JEOL JSM-7800F / Thermo Fisher Apreo 2 / Zeiss GeminiSEM 560

Accelerating Voltage Kv

10

Beam Current Na

0.54

Working Distance Mm

10

Pixel Size Nm

7.1

Magnification

20,000x

Detector

Everhart-Thornley (SE2) + in-lens (SE1)

Image Size

1024x768

Signal Chain Diagram

Experimental setup diagram for Scanning Electron Microscopy

Key References

  • Goldstein et al., 'Scanning Electron Microscopy and X-ray Microanalysis', Springer (2018)

Canonical Datasets

  • SEM Dataset for Nanomaterial Segmentation (Aversa et al.)
  • NIST SEM calibration images

Related Modalities

TEM STEM Electron Tomo 4D-STEM Electron Holography EBSD EELS Cryo-Electron Tomography (Cryo-ET)

Benchmark Pages

SEM