Quantum Simulation Lab (QSimLab)
QSimLab is an integrated, research-grade environment for building, simulating, postprocessing and visualizing materials and electronic devices – from atomic structure to experimentally relevant observables.
The platform connects crystal and device construction, VR/AR-enabled visualization, multi-terminal transport (NEGF), bulk linear and nonlinear response, and high-performance distributed computing through a [soon to be integrated] custom-built job distribution engine.
End-to-End Capabilities
Crystal & Device Builder
Create and modify atomic structures and devices directly in the browser. Import from CIF/POSCAR and other common formats, or construct multi-terminal junctions using the interactive geometry panel and export into various 2d/3d formats.
Tools include unit-cell conversion, supercell generation, symmetry-aware transformations, interface matching, and atom-level editing with periodic-table based species selection.
Interactive 3D & VR/AR Visualization
Explore structures and devices in real-time 3D using WebGL and WebXR. Toggle lattice vectors, bonds, magnetic moments, and device regions; export publication-quality snapshots and 3D models.
When supported, the same scenes can be entered in VR/AR for immersive inspection of complex interfaces, heterostructures, and transport geometries.
Multi-Terminal Quantum Transport (NEGF)
Build and analyze multi-terminal devices through a dedicated NEGF environment. Compute lead properties, band structures, and transmission spectra under realistic boundary conditions using tight-binding or OpenMX Hamiltonians.
Multi-terminal configurations and bias profiles are handled in a unified interface suitable for both exploratory studies and systematic device design.
Linear & Nonlinear Response
Access bulk linear-response modules for conductivity and optical response, with an emphasis on frequency-dependent and tensorial properties.
Nonlinear extensions (e.g. higher-order optical response and light–matter interaction tools) are being developed to link microscopic electronic structure with measurable observables.
Data Visualization & Analysis
Visualize results through interactive plots and 3D fields: Fermi-surface cuts, Brillouin-zone paths, spatial distributions, volumetric data, and device-resolved quantities.
All visualization tools are designed with reproducibility in mind, allowing you to reload stored sessions and regenerate figures consistently across runs.
SpeedFarm Distributed Computing
Offload heavy simulations to the SpeedFarm distributed job engine. SpeedFarm orchestrates jobs across local and remote resources, tracking usage and enabling scalable studies that go beyond a single workstation.
Education & Classroom Use
The platform includes a growing set of theoretical notes and hands-on tools (see the Notes tab) designed for introductory to advanced courses in condensed-matter physics, materials science, and nanoelectronics.
Instructors can use these modules for live demonstrations, while students can directly manipulate structures, run simplified simulations, and visualize the underlying physics.
Ongoing and Upcoming Modules
QSimLab is actively evolving (Currently about 20% implemented). The following capabilities are under development and will be integrated into the same workflow:
- Hartree–Fock+NEGF solvers.
- Density Functional Theory (DFT) interfaces and workflows for electronic structure, forces, and derived observables.
- Magnetic properties: magneto-crystalline anisotropy, interatomic exchange tensors, and Gilbert damping extracted from ab-initio or tight-binding models.
- Maxwell's solvers: Use calculated optical response in equilibrium and under finite DC bias, to simulate electromagnetic propagation by solving Maxwell’s equations.
These modules are intended to connect seamlessly with existing device and bulk workflows, so that the same structures and Hamiltonians can be used across transport, spectroscopy, and magnetism.
Device Setup
Loaded Structures
Isotopes:
Electronic Structure Calculations:
Projects Discussion Panel (Conetents stay on server temporarily.)
About QSimLab
QSimLab is designed as a bridge between interactive teaching tools and cutting-edge research infrastructure. The same interface that allows a student to explore a crystal structure for the first time can be used by researchers to prototype realistic device geometries, launch large-scale simulations, and prepare publication-quality figures.
The platform emphasizes transparent, reproducible workflows. Structures, Hamiltonians, simulation parameters, and visualizations are organized so that results can be revisited, shared, and extended without rebuilding the entire pipeline. Community-oriented features enable users to discuss projects, compare methodologies, and refine theoretical interpretations.
For the classroom, the Notes section and associated tools will provide guided, hands-on introductions to topics such as fundamental quantum methonics, band structure, Brillouin-zone sampling, transport in mesoscopic devices, and optical response. For advanced users, the same environment scales to multi-terminal NEGF, nonlinear response, and forthcoming Hartree–Fock, DFT, and magnetic-property modules, all accessible through a single, coherent front end.
Getting Started
- Build or load a structure in the Geometry tab: upload CIF/POSCAR files or construct devices and terminals directly.
- Choose a simulation mode: use the Electronic Structure and Post Processing tabs for bulk response, or the Multi-Terminal Device interface for NEGF-based transport.
- Explore and visualize results: use the Data Visualization tools and VR/AR-enabled viewers to analyze structures, fields, and spectra, and export figures or 3D assets for publications and presentations.
- Scale up with SpeedFarm when available, to distribute demanding calculations over multiple machines while keeping the same browser-based control panel.
Cite this website
@misc{QSimLab,
title = {Quantum Simulation Lab},
howpublished = {\url{https://www.qsimlab.com/}},
note = {Accessed 2025-12-05},
year = {2025}
}
Latest Updates
- Kubo / Linear Response: fixed a geometry/orbital parsing issue in the Kubo-based conductivity workflow that could produce unphysical tensor components (e.g., spurious symmetric off-diagonal terms) when building velocity operators from ab-initio models.
-
NEGF: current-induced force / derivative operators: added local derivative observables
based on finite differences of the Hamiltonian,
dH/dv(anddH/dv1,dH/dv2, … when multiple derivative tokens are used). This supports force-like quantities of the formF = -⟨∂H/∂q⟩once the user maps the parametervto a structural coordinate or generalized mode.- Derivative operators are built as
(H(v+dv)-H(v-dv))/(2dv). - The finite-difference step is user-controlled (e.g.
obs_dv_fd). - These operators can be included in the same observable/correlation pipeline as velocity operators.
- Derivative operators are built as
-
NEGF: Floquet extension (time-periodic drive): added a Floquet mode that extends the device
Hamiltonian into harmonic space (user-selected number of harmonics and drive frequency).
- Supports harmonic-resolved transmission and harmonic-resolved observables.
- Includes a vs-ω (frequency sweep) mode at fixed energy
E0. - Optional “central-only” Floquet: keep leads static while extending only the central region.
-
NEGF: non-equilibrium cross-correlation: added an optional cross-correlation module for pairs
of operators (e.g. velocities and/or
dH/dv*) computed from lead-resolved non-equilibrium density-matrix/Green-function objects:CC ∼ O1 · ρa · O2 · ρb.- Can be evaluated at a single ω* and/or computed vs ω on a user-defined grid.
- Uses the same operator-selection infrastructure as observables.
-
Distributed execution (SpeedFarm): integrated SpeedFarm as an execution engine for both NEGF
and Kubo workflows, enabling distributed job submission/collection for parameter sweeps and heavy runs [Needs more testing].
- UI-selectable compute engine (local Python / Pyodide / SpeedFarm).
- Designed to split and merge work across k-points and sweep dimensions.
-
Large-system Hamiltonian handling: improved ab-initio Hamiltonian parsing and storage with
an atom-pair block format (
atom_pairs_v1) to reduce memory pressure and improve load/assembly time.- Stores Hamiltonian/overlap blocks per atomic pair instead of dense full matrices.
- Includes a compact “block library + references” design to deduplicate repeated blocks.
- Intended for large OpenMX supercells and similar ab-initio exports.
-
Tunnel-junction supercell detection (PBC → NEGF device): added a module that detects
“embedded lead segments” inside large ab-initio supercells (common in tunnel-junction/slab setups) and
automatically identifies outer lead layers vs central/junction region so the ends can be treated as
semi-infinite leads.
- Conservative detector: triggers only when both ends show a clear periodic lead signature.
- Supports principal-layer thickening when hoppings extend beyond nearest neighbor cells.
Electronic Structure Viewer
Post-Processing Tools
Data Visualization Tools
Interactive charts and visualization tools for simulation results.
SpeedFarm Job Engine
This tab embeds the SpeedFarm distributed job engine UI. Use it to monitor workers, submit test tasks, and debug distributed runs driven from the QSimLab environment.