A Python/PySide6 desktop application for exploring transmission electron microscopy, electron optics and simulated detector signals.
The simulator provides an offline environment for learning instrument controls, exploring parameter responses and developing experimental workflows. This can help researchers make better use of precious instrument time, by preparing and testing ideas before using a microscope. It also provides a foundation for future AI for experiments, including AI-guided tuning and experiment design. The simulator does not currently provide autonomous microscope control.
- Editable microscope geometry and hardware settings, with electron emission, extraction, acceleration, lenses, deflectors, stigmators and apertures.
- Interactive ray diagrams, transverse beam plots, flight-time colouring, magnetic-field views and individually adjustable virtual electron paths.
- A separate coherent-beam development page with explicit tip-boundary inputs, cached exact-Z observations and clear unsupported-operator limits.
- Hardware tuning with beam measurements and comparison against a saved baseline.
- Specimen scattering, detector signals, STEM scanning, EDS and optional energy-filter modelling.
- Live tuning with user-selected calculation endpoints, cached upstream states and resumable particle calculations.
- Compressed result export/import (
.temresult), reusable startup results and separate virtual-electron sessions (.temdiag).
Classical particle transport remains the default. Independent coherent development resumed on 2026-10-02; the default tip-to-image chain and arbitrary Z within the gun or material are not yet qualified. See coherent-beam inputs and limits. Models target physical mechanisms and qualitative parameter trends, rather than a calibrated commercial instrument or a fully validated microscope.
Use Run high-accuracy once to update Ray Diagram. Then set specimen or detector parameters and click Calculate on that page; compatible executed beam states before the specimen are reused, and downstream rays update with the result. Tab changes do not calculate. Live tuning retains its separate cutoff and continuation controls.
For fast CIF-dependent HAADF/BF/DF images, choose Projected atoms (fast approximation) in Scanning Image, set the scan pixel size and click Calculate. This non-wave model combines actual CIF positions/occupancies, the executed probe footprint and physical downstream detector acceptance. It estimates thin-sample independent-atom contrast; BF/DF do not include Bragg interference or channeling. A broad probe or coarse sampling still removes atomic detail. Material particle paths retains the finite-volume particle transport readout. Auto contrast changes display limits only; actual intensity ranges remain visible. Loading a CIF does not enable coherent imaging.
The Sample page starts with Vacuum sample. To add material, import your own CIF / MCIF, set its dimensions and orientation, then click Calculate Sample. This calculates specimen interactions and updates downstream rays; STEM images and EDS spectra use their own page buttons. No built-in reference CIF library or automatic material selection is provided.
On Windows, install 64-bit Python 3.12 (recommended), then run these commands from the project folder:
py -3.12 setup_env.py
.venv\Scripts\python.exe main.pySetup creates or reuses .venv, installs dependencies and checks the environment.
Python 3.11–3.13 is supported; the interface uses PySide6 6.8.3. Internet access
is normally needed for missing dependencies. Run py -3.12 setup_env.py --help
for offline installation and optional GPU settings.
For subsequent launches, run only the second command. In PyCharm, select
.venv\Scripts\python.exe as the interpreter and run main.py.
Project map / 项目地图: files, features, UI, tests and refactoring candidates.