Electron Energy Loss Spectroscopy

eels Electron Microscopy Spectroscopic Wave Optics
View Benchmarks (1)

STEM-EELS measures the energy distribution of electrons transmitted through a thin specimen, where inelastic scattering events encode information about elemental composition, bonding, and electronic structure. The energy loss spectrum contains core-loss edges (characteristic of specific elements) and low-loss features (plasmons, band gaps). A magnetic prism spectrometer disperses the energy spectrum onto a position-sensitive detector. Spectrum imaging acquires a full spectrum at each scan position, enabling elemental mapping with atomic-scale spatial resolution.

Forward Model

Energy Loss Cross Section

Noise Model

Poisson

Default Solver

fourier ratio

Sensor

SCINTILLATOR_CCD

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 Lambda energy Energy Disperser D g, η₁ CCD Spectrometer
Spec Notation

P(e⁻) → Λ(energy) → D(g, η₁)

Benchmark Variants & Leaderboards

EELS

Electron Energy Loss Spectroscopy

Full Benchmark Page →
Spec Notation

P(e⁻) → Λ(energy) → D(g, η₁)

Standard Leaderboard (Top 10)

# Method Score PSNR (dB) SSIM Trust Source
🥇 DiffEELS 0.882 39.3 0.954 ✓ Certified Gao et al. 2024
🥈 PhysEELS 0.853 37.9 0.942 ✓ Certified Chen et al. 2024
🥉 SwinEELS 0.828 36.7 0.932 ✓ Certified Wang et al. 2023
4 TransEELS 0.792 35.1 0.915 ✓ Certified Li et al. 2022
5 N2V-EELS 0.731 32.6 0.876 ✓ Certified Krull et al. 2019
6 DnCNN-EELS 0.669 30.0 0.838 ✓ Certified Kovarik et al. 2016
7 ICA-EELS 0.595 27.1 0.786 ✓ Certified Bosman et al. 2006
8 MLS-EELS 0.530 24.5 0.744 ✓ Certified Verbeeck & Van Aert 2004
9 PowerLaw-EELS 0.463 21.8 0.699 ✓ Certified Egerton 2011
Mismatch Parameters (3) click to expand
Name Symbol Description Nominal Perturbed
energy_dispersion ΔD_E Energy dispersion error (eV/channel) 0 0.002
zero_loss_shift ΔE_0 Zero-loss peak shift (eV) 0 0.3
aberration ΔC_c Chromatic aberration error (%) 0 2.0

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: energy loss cross section — Mismatch modes: plural scattering, channel to channel gain, drift during acquisition, radiation damage

G2 — Noise Characterization Is the noise model correctly specified?

Noise: poisson — Typical SNR: 3.0–20.0 dB

G3 — Calibration Quality Are instrument parameters accurately measured?

Requires: energy dispersion, zero loss alignment, collection angle, convergence angle

Modality Deep Dive

Principle

Electron Energy Loss Spectroscopy measures the energy lost by transmitted electrons due to inelastic interactions with the specimen. The energy-loss spectrum contains characteristic edges corresponding to inner-shell ionization of specific elements, enabling elemental mapping with atomic spatial resolution. Near-edge fine structure (ELNES) reveals chemical bonding, and low-loss features probe band structure and optical properties.

How to Build the System

Attach a post-column energy filter (Gatan GIF Quantum/Continuum) to a TEM/STEM. For STEM-EELS spectrum imaging: scan the probe and record a full energy-loss spectrum (0-2000 eV range) at each pixel. Use a monochromated source (ΔE < 0.3 eV) for near-edge fine structure studies. Energy dispersion is typically 0.1-0.5 eV/channel. Acquire both core-loss edges (elemental maps) and low-loss region (thickness mapping, optical properties).

Common Reconstruction Algorithms

  • Background subtraction (power-law fitting before edge onset)
  • Multiple linear least-squares (MLLS) fitting for overlapping edges
  • Principal component analysis (PCA) for denoising spectrum images
  • Kramers-Kronig analysis for optical constants from low-loss EELS
  • Deep-learning EELS denoising and quantification

Common Mistakes

  • Specimen too thick causing plural scattering that distorts edge shapes
  • Incorrect background model for edge extraction (wrong fitting window)
  • Energy drift during long spectrum-image acquisitions
  • Not accounting for plural scattering when quantifying elemental ratios
  • Beam damage altering the specimen chemistry during EELS acquisition

How to Avoid Mistakes

  • Keep specimen thickness < 0.5 inelastic mean free path (t/λ < 0.5)
  • Fit background in a window just before the edge; use multiple-window methods if needed
  • Apply energy drift correction using the zero-loss peak or a known edge
  • Deconvolve plural scattering using Fourier-log method before quantification
  • Use low-dose protocols and fast spectrum imaging to minimize beam damage

Forward-Model Mismatch Cases

  • The widefield fallback produces a 2D spatial image, but EELS acquires energy-loss spectra at each probe position — the spectral dimension encoding elemental composition (core-loss edges) and electronic structure (near-edge fine structure) is entirely absent
  • Each EELS spectrum contains characteristic ionization edges (e.g., C-K at 284 eV, O-K at 532 eV) that identify elements with atomic spatial resolution — the widefield spatial blur cannot access spectroscopic chemical information

How to Correct the Mismatch

  • Use the EELS operator that models energy-loss spectrum formation: each probe position produces a spectrum with background (power-law), core-loss edges (proportional to elemental concentration), and near-edge fine structure (bonding information)
  • Quantify elemental maps using background subtraction and edge integration, or MLLS fitting for overlapping edges; apply PCA denoising to spectrum images before quantification

Experimental Setup

Instrument

Gatan Quantum GIF / Gatan Continuum / Nion HERMES

Accelerating Voltage Kv

100

Energy Resolution Ev

0.3

Dispersion Ev Per Ch

0.1

Collection Angle Mrad

30

Dwell Time Ms

50

Spectrum Range

core-loss + low-loss

Analysis

elemental mapping, ELNES fine structure

Signal Chain Diagram

Experimental setup diagram for Electron Energy Loss Spectroscopy

Key References

  • Egerton, 'Electron Energy-Loss Spectroscopy in the Electron Microscope', Springer (2011)

Canonical Datasets

  • EELS Atlas (Ahn & Krivanek)

Benchmark Pages