Fermi — Semiconductor Detector Simulator for Photon-Counting CT

Fermi couples a microscopic Fermi-band / carrier-transport model with a macroscopic X-ray response model to evaluate semiconductor detector materials for Photon-Counting CT (PCCT) workloads in the 40–190 keV energy band.

---

Overview

Semiconductor PCCT detectors must simultaneously deliver high absorption efficiency, high count-rate capability, and good energy resolution under a harsh high-flux X-ray load. Fermi lets you explore these trade-offs for a whole library of candidate materials inside a single, interactive desktop GUI.

The simulator combines two physics layers:

• Fermi-band / transport layer — intrinsic carrier concentration, Fermi level, resistivity, dark (shot) current and its contribution to electronic noise, plus the Hecht charge-collection efficiency and the space-charge / polarization field distortion.
• X-ray response layer — real attenuation from the NIST Xraylib database, depth-dependent photon interaction, K-edge / escape effects, and Monte-Carlo spectral generation.

The result is a physically-grounded view of spectral response, noise, CCE, polarization, attenuation and high-flux performance that can be compared side-by-side across materials.

Scope note — Fermi uses a 1-D geometry model. It reproduces the relative physical laws and differences between materials very well, but absolute values (e.g. final energy resolution) should be treated as theoretical references under ideal conditions rather than engineering measurements of a specific commercial detector.

---

Features

• Interactive desktop GUI (PyQt6) with side-by-side multi-material comparison (up to 4 materials at once), English / 中文 switching, and a custom-material repository importer.
• 7 analysis panels per material: Spectral Response, Band Structure, Charge Collection & Internal Field, Polarization Effect, X-ray Attenuation, Performance vs Temperature, and Performance vs Photon Flux.
• PCCT Comparison Summary with a sortable evaluation table and a 6-axis PCCT Suitability Radar (Efficiency, Speed, Resolution, Stability, High-Flux, Peak fidelity).
• Real cross-section data via xraylib (with a graceful analytic fallback when unavailable).
• CLI entry point for scripted batch simulation and CSV/JSON export.
• Cross-platform packaging through GitHub Actions (Windows .exe + macOS .dmg).

---

Screenshots

Detailed results & spectral response

Two materials compared at 120 keV — note the per-material statistics header (efficiency, noise FWHM, dark current, ENC, resistivity) and the K-edge structure in the spectrum.

PCCT comparison summary & suitability radar

Chinese UI

All analysis panels (click to expand)

Panel	Plot
Spectrum
Band structure (Fermi level)
Charge collection & internal field
Polarization effect
X-ray attenuation (K-edges)
Performance vs temperature
Performance vs photon flux

---

Physics Models

Module	Quantity	Model
Fermi band	Intrinsic carriers	n_i = sqrt(N_c·N_v)·exp(−E_g/2kT)
Fermi band	Fermi level	E_i = E_g/2 + (kT/2)·ln(N_v/N_c)
Transport	Resistivity	ρ = 1 / (q·n_i·(μ_e + μ_h))
Transport	Dark-current shot noise	σ = sqrt(2·q·I_dark·τ_int) → ENC → FWHM
Transport	Charge collection	Hecht equation with drift lengths λ = μτ·E
Transport	Polarization	space-charge field distortion vs. time / flux
X-ray	Attenuation	NIST Xraylib CS_Total_CP (K-edge aware)
X-ray	Absorption efficiency	1 − exp(−μ·d) with depth-dependent interaction
X-ray	Spectral response	Monte-Carlo photopeak + Fano-limited Gaussian broadening

The complete design write-up is in doc/design.md; the underlying scientific background is collected in the kb/ knowledge base (Fermi integration, Ge & semiconductor detectors, 40–190 keV band).

---

Materials Library

Fermi ships with 9 built-in materials and can import more from a remote repository.

Material	E_g (eV)	W (eV)	Density (g/cm³)	μτ(e) / μτ(h)
Si	1.12	3.6	2.33	high / high
Ge	0.66	2.96	5.32	high / high
CdTe	1.44	4.43	5.85	3.3e-3 / 1.0e-4
CZT	1.57	4.6	5.78	5.0e-3 / 4.0e-5
TlBr	2.68	6.5	7.56	3.0e-4 / 4.0e-6
HgI2	2.13	4.2	6.40	1.0e-4 / 4.0e-5
GaAs	1.42	4.2	5.32	8.5e-2 / 4.0e-4
Perovskite (MAPbI3)	1.55	5.0	4.16	6.0e-5 / 6.0e-5
4H-SiC	3.26	7.8	3.21	9.5e-4 / 1.2e-4

Additional importable materials (GaAs, InP, Diamond, MAPbI3, HgI2) live in data/remote_repo.json.

---

Installation

git clone https://github.com/hg3992260/Fermi_PCCT.git
cd Fermi_PCCT

# create / activate an environment (example with conda)
conda create -n Fermi python=3.11 -y
conda activate Fermi
pip install -r requirements.txt

Dependencies (requirements.txt): numpy, matplotlib, PyQt6, xraylib.

Run the GUI

python src/gui.py
# or, from the project root
conda run -n Fermi python I:\Fermi\src\gui.py

Run from the command line

python -m src.main --material CZT --energy 120 --temp 300 --bias 500 --thick 2.0 --output my_run

This writes data/my_run_spectrum.csv and data/my_run_stats.json.

---

Building Desktop Apps

GitHub Actions build native binaries on every push or manual dispatch:

Platform	Workflow	Artifact
Windows	.github/workflows/windows-build.yml	Fermi-windows → Fermi.exe
macOS	.github/workflows/macos-dmg.yml	Fermi-macos-dmg → Fermi.dmg

Download: repository Actions tab → select a run → Artifacts.

macOS Gatekeeper note — the DMG/App is not code-signed or notarized. Right-click → Open to bypass, or run:

xattr -dr com.apple.quarantine /Applications/Fermi.app

Icons

• Runtime window/taskbar icon: resources/Atom.jpeg (loaded in-app)
• Executable icon: resources/Atom.ico (PyInstaller --icon)

Regenerate the .ico with:

python tools/make_ico.py

---

Analysis Tools

The tools/ directory contains standalone report generators that reuse the same core physics:

Script	Purpose
pcct_material_report_calc.py	Rank all materials by a composite PCCT score across pixel pitches
pcct_material_report_render.py	Render the JSON report into Markdown/HTML
pcct_polarization_compare.py	Tabulate ΔFWHM flux-degradation across materials
plot_polarization_curves.py	Polarization field-strength curves (CdTe/CZT/Si/GaAs)
make_ico.py	Build the application icon

---

Repository Structure

Fermi/
├── src/
│   ├── gui.py            # PyQt6 desktop application (main window, plots, i18n)
│   ├── simulator.py      # Detector + X-ray response integration (MC, Hecht, flux)
│   ├── physics.py        # Fermi-Dirac, resistivity, Hecht, K-edges, fallback xraylib
│   ├── material.py       # Material model + manager (load/save/import)
│   ├── monte_carlo.py    # Photon transport Monte-Carlo
│   ├── field_solver.py   # Internal electric-field / space-charge solver
│   ├── translations.py   # English / Chinese UI strings
│   └── main.py           # CLI entry point
├── data/                 # Material library, example spectra, remote repo
├── kb/                   # Scientific background knowledge base
├── doc/design.md         # Design document
├── tools/                # Standalone analysis & reporting scripts
├── resources/            # Icons (Atom.ico / Atom.jpeg)
├── docs/screenshots/     # README screenshots
└── .github/workflows/    # Windows + macOS packaging

---

Credits

Developed by Christ — Christ.paul90@gmail.com.

Physics background adapted from the kb/ knowledge base; X-ray cross-sections from the NIST Xraylib project.

License

Released under the MIT License. © 2026 Christ (hg3992260).