STEM HAADF & EELS Atomic Metrology Lab

Model Aberration-Corrected Scanning Transmission Electron Microscopy (STEM), High-Angle Annular Dark-Field ($Z$-contrast, $I \propto Z^{1.75}$) Rutherford scattering, core-loss Electron Energy Loss Spectroscopy (EELS), sub-Ångström probe formation, and 20-year cleanroom fab analytical lab DCF economics.

Aberration-Corrected STEM Atomic Z-Contrast Milestone 900
Electron Probe Size
0.58 Å
Sub-0.5 Å (Cs-Corrected)
de Broglie Wavelength
2.51 pm
Relativistic λ at 200 kV
Specimen Thickness (t/λ)
0.31
Low-loss plasmon ratio
Relativistic Velocity
69.5% c
Beam velocity fraction
Inelastic Mean Free Path
80.0 nm
λ_inelastic in silicon
20-Yr Tool Suite NPV
+$45.62M
IRR 48.2% | 1.8 Years
FIB Lamella Throughput
1,200 / Year
4 lamellas/day capacity
FA Turnaround Time
< 6 Hours
vs. 5-7 days external lab

STEM & Optics Controls

200 kV
60 kV (2D Gentle) 300 kV (Penetration)
Cancels spherical aberration, enabling sub-0.5 Å probe
25 mrad
25 nm
5 nm (Ultra-Thin) 80 nm (Plural Scatter)
68 mrad
40 mrad (LAADF / Strain) 100 mrad (Pure Z)
15 μs

Atomic-Scale Visualizer

HAADF-STEM Z-Contrast & Simulated Atomic Columns

Top: 1D horizontal intensity profile reflecting Rutherford scattering cross-section ($I \propto Z^{1.75} \cdot t$). Bottom: simulated 2D atomic column cross-section showing sub-Ångström structural mapping. High-$Z$ atoms ($\text{W } Z=74$, $\text{Hf } Z=72$) scatter brightly, directly distinguishing metal gate electrodes from low-$Z$ silicon channels.

Core-Loss Electron Energy Loss Spectrum (EELS)

Energy-loss spectrum demonstrating elemental ionization edges: $\text{Si } L_{2,3}$ ($99\text{ eV}$), $\text{N } K$ ($401\text{ eV}$), $\text{Ti } L_{2,3}$ ($456\text{ eV}$ & $462\text{ eV}$ white-line splitting), $\text{O } K$ ($532\text{ eV}$), and $\text{Hf } M_{4,5}$ ($1660\text{ eV}$) superimposed on power-law background ($A E^{-r}$).

Atomic-Scale Chemical Composition Line-Scan

Quantitative atomic concentration profiles across the dielectric/channel interface, tracking equivalent oxide thickness (EOT) scaling and detecting oxygen scavenger migration or interfacial sub-oxide interdiffusion down to sub-nanometer precision.

20-Year Cleanroom Analytical STEM Suite DCF Financial Model

Evaluate the capital investment of commissioning an in-house analytical electron microscopy facility (\$6.5M - \$9.5M for Dual-Beam FIB-SEM lamella prep + Aberration-Corrected STEM) versus outsourcing failure analysis to third-party labs (\$3,500/lamella with 5 to 7 days turnaround). Rapid in-house diagnostics catch process drift early, preventing multi-million dollar wafer lot scrap.

Timeline CapEx Outsourcing & Yield Benefit Suite OpEx & Maintenance Net Annual Cash Flow Cumulative Discounted Cash Flow

Theory & Physics of Aberration-Corrected STEM & EELS

1. Rutherford Scattering & High-Angle Annular Dark-Field

At large scattering angles ($\theta > 50\text{ mrad}$), coherent Bragg diffraction is washed out by thermal diffuse scattering. The differential cross-section approaches atomic nucleus Rutherford scattering:

dσ/dΩ = (Z * e^2 / (16 * π * ε_0 * E_0))^2 * (1 / sin^4(θ/2))

Integrating over annular detector geometry yields $I_{\text{HAADF}} \propto t \sum Z_i^\alpha$, where $\alpha \approx 1.7\text{–}1.9$. This powers direct atomic number identification without phase-contrast contrast inversions.

2. Hardware Cs Aberration Correction

Round magnetic lenses inevitably possess positive spherical aberration ($C_s > 0$). Incorporating multipole corrector elements introduces negative spherical aberration:

d_probe ≈ 0.61 * λ / α_conv (diffraction limit)

Cancelling $C_s$ allows increasing the convergence semi-angle $\alpha_{\text{conv}}$ from $10\text{ mrad}$ to over $25\text{–}30\text{ mrad}$, compressing probe diameter below $0.5\text{ Å}$ ($50\text{ pm}$) for true atomic column resolution.

3. Core-Loss EELS & Chemical Bond Fingerprinting

Fast beam electrons undergo inelastic Coulombic collisions that eject core-level electrons into unoccupied conduction band states:

N = [ I_edge(β, Δ) / I_0(β) ] * [ 1 / σ_k(β, Δ) ]

The energy-loss near-edge structure (ELNES) maps the local density of unoccupied states (LDOS), resolving crystal-field splitting (e.g. $\text{Ti } L_3/L_2$ white lines) and metal work function work states.

4. Specimen Thickness & Inelastic Mean Free Path

The log-ratio technique quantifies local lamella thickness ($t$) relative to the total inelastic mean free path ($\lambda$):

t / λ = ln( I_total / I_zero_loss )

Maintaining $t/\lambda < 0.3\text{–}0.5$ via precision low-kV focused ion beam (FIB) final polishing is essential to avoid plural scattering artifacts in high-resolution core-loss spectroscopy.

Formative Assessment: STEM HAADF & EELS Metrology

Question 1: In HAADF-STEM imaging of a 2nm GAA nanosheet stack, why does the tungsten gate fill appear dramatically brighter than the silicon nanosheet channel?

Question 2: What is the main reason an in-house Cs-corrected STEM lab delivers rapid financial payback in semiconductor manufacturing?

Question 3: In EELS analysis, what is the consequence of analyzing a TEM lamella that is too thick (t/λ > 1.0)?

Frequently Asked Questions

In Scanning Transmission Electron Microscopy (STEM), High-Angle Annular Dark-Field (HAADF) detectors collect electrons scattered to high angles (typically greater than 50 mrad). At these high collection angles, coherent Bragg diffraction is largely suppressed, and thermal diffuse scattering dominates. The scattering cross-section approaches classical Rutherford scattering, causing atomic column intensity to scale roughly with atomic number Z squared (I proportional to Z^1.7 to Z^1.9). Consequently, heavier elements like tungsten (Z=74) and hafnium (Z=72) appear dramatically brighter than silicon (Z=14) or oxygen (Z=8).

Uncorrected electromagnetic round lenses inherently suffer from positive spherical aberration (Cs), where marginal electron rays refract more strongly than paraxial rays, causing severe probe broadening (d_probe approximately 1.2 to 2.0 Angstroms). Hardware Cs correctors introduce non-round multipole lenses (hexapoles or quadrupoles/octupoles) that generate negative spherical aberration, cancelling the lens defects. This enables larger probe semi-convergence angles (25 to 35 mrad) and achieves sub-0.5 Angstrom atomic probe diameters.

Electron Energy Loss Spectroscopy (EELS) analyzes the energy lost by beam electrons via inelastic interactions with the specimen. In the core-loss regime (above 50 eV), EELS reveals sharp ionization edges characteristic of specific elements (such as the Si L-edge at 99 eV, Ti L-edge at 456 eV, O K-edge at 532 eV, and Hf M-edge at 1660 eV). Furthermore, energy-loss near-edge structure (ELNES) provides oxidation state and chemical bonding details, resolving sub-nanometer interdiffusion across gate dielectrics.

As electron beam electrons traverse a TEM lamella, the probability of multiple inelastic scattering events increases exponentially with thickness. The relative thickness parameter t/lambda (where lambda is the inelastic mean free path, approximately 80 nm in silicon) is extracted from the low-loss spectrum using the log-ratio technique. In lamellas where t/lambda exceeds 0.5, plural scattering distorts core-loss edge shapes, broadens probe channeling, and degrades spatial resolution, requiring precise FIB low-kV milling down to 20 to 30 nm.

An advanced analytical STEM suite costs between $6.5M and $9.5M in CapEx (including Dual-Beam FIB lamella preparation and Cs-corrected STEM hardware), but replaces high third-party failure analysis costs ($3,500 per sample) while cutting analysis turnaround time from 5 to 7 days down to under 6 hours. Fast turnaround physical defect verification prevents multi-million dollar wafer lot scrap excursions caused by gate oxide bridging or nanosheet release defects, typically delivering an IRR exceeding 36% and a discounted payback period under 2.5 years.

Related Semiconductor & Metrology Simulations

EBSD Crystallography Lab

Model Kikuchi electron diffraction bands, CSL twin boundaries, GND dislocation density, and electromigration MTTF.

Raman Spectroscopy Lab

Model Raman spectroscopy, phonon strain tensors, Stokes and Anti-Stokes shifts, 2D materials, and fab metrology DCF.

SIMS Depth Profiling Lab

Model Secondary Ion Mass Spectrometry (SIMS) depth profiling, dopant concentration, knock-on mixing, junction depth, and fab metrology DCF.

KPFM Surface Potential Lab

Model Kelvin Probe Force Microscopy (KPFM), contact potential difference, nanoscale work function, band bending, and fab metrology DCF.

TXRF Wafer Metrology Lab

Model Total Reflection X-Ray Fluorescence (TXRF), trace metal surface contamination, standing waves, and fab DCF.

Picosecond Laser Ultrasonics Lab

Model picosecond laser ultrasonics, acoustic impedance, echo transit times, opaque metal film thickness, and fab DCF.

HRXRD & Epitaxial Strain Lab

Model dynamical X-ray diffraction, reciprocal space mapping (RSM), strain relaxation, and fab metrology DCF.

All Business Simulations & Worksheets →

Browse our comprehensive directory of over 900 engineering, physics, financial, and manufacturing simulation labs.