Computational Modelling
Of
Semiconductors
Nanosystems Advisory is building NSDW (Nanosystems Defect Workflows) a reproducible computational platform for semiconductor defect modelling across high-performance computing systems.
NSDW standardises the path from material structure and simulation methodology to HPC execution, provenance, validated results and, ultimately, defect intelligence.
Automate. Validate. Scale.
automate
validate
scale
extend
Turn repeatable computational methods into structured workflows rather than one-off simulation scripts.
Track methodology, convergence, execution and results with reproducible provenance.
Run the same scientific workflow across local and national HPC infrastructure.
Build towards automated defect campaigns, machine-learning acceleration and a structured semiconductor defect knowledge platform.
THE PLATFORM
NSDW — Nanosystems Defect Workflows
From structure to validated simulation
NSDW is the computational engine being developed by Nanosystems Advisory.
The current platform integrates semiconductor structure handling, simulation methodology, CP2K execution, convergence workflows, reporting, AiiDA provenance and HPC execution into a single reproducible framework.
Already validated through real first-principles calculations on UK and European HPC infrastructure.
Current development focus: multi-HPC portability
structure
methodology
CP2K execution
convergence
validated results
Defect intelligence
CAPABILITIES
Built for computational materials science at scale
01
Workflow automation
Standardised project structures, persisted methodologies and reusable production recipes.
02
HPC execution
First-principles workloads submitted and executed through structured HPC interfaces rather than machine-specific manual workflows.
03
Provenance
AiiDA-backed tracking connects input structures, methods, calculations and results.
04
Analysis
Automated parsing, convergence assessment and machine-readable JSON/CSV outputs.
ROADMAP
Development Roadmap
NOW
HPC Workflow Infrastructure
CP2K execution
methodology persistence
Production recipes
Convergence workflows
AiiDA provenance
ARCHER2 validation
LUMI-C validation
NEXT
Automated defect campaigns
Generic HPC profiles
Cirrus integration
Defect structure generation
Campaign orchestration
Multi-HPC execution
ML-assisted screening
THEN
Defect physics
Charged defects
Formation energies
Charge-transition levels
Migration barriers
NEB workflows
Defect concentration
LONG TERM
Semiconductor defect intelligence
Structured defect datasets
Cross-material comparison
Searchable defect knowledge
Active-learning workflows
Predictive materials screening
THE PLATFORM
Compute where the science belongs
NSDW is being designed around portable scientific workflows rather than a single supercomputer.
Current and planned infrastructure includes European Tier-0/Tier-1 and UK Tier-2 systems, allowing workflows to move between machines while retaining the same methodology and provenance.
VALIDATED
LUMI-C
Validated AiiDA-to-CP2K execution.
VALIDATED
ARCHER2
Real NSDW-to-Slurm-to-CP2K workflow validation.
NEXT
Cirrus
Next platform in the portability roadmap.
PLANNED
GPU Infrastructure
Planned for machine-learning potential training and defect-screening workflows.
Scientific software should be reproducible by design
PRINCIPLES
Semiconductor defect calculations often depend on complex combinations of structures, basis sets, convergence parameters, functionals, charge states and HPC configurations.
NSDW is being built so those choices are represented explicitly, versioned and reproducible allowing computational campaigns to scale without losing scientific traceability.
InP
zinc-blende · III-V
zinc-blende · III-V
GaSb
GaN
wurtzite · III-nitride
zinc-blende · III-V
GaAs
IGZO
amorphous oxide
GaN
wurtzite · III-nitride
The same method, run across five different lattice types
Nitride, oxide, arsenide, antimonide, phosphide. Different bonding, different structures, one repeatable pipeline. See the full method on the Services page.
Proven on
Get in Touch
info@nanosystemsadvisory.com
+44 (075) 35506520
London, United Kingdom