Sequential Doping Workflow
The sequential doping workflow is designed for gradual composition-by-composition doping studies. Instead of generating each target concentration independently from the pristine host, the workflow uses the lowest-energy relaxed structure from the previous composition as the base structure for the next composition.
Command
dopingflow sequential-run -c input.toml
Execution Logic
For each composition generated from the [doping] section, the workflow creates
a separate step folder under [sequential].outdir:
sequential_structures/
step_001_Sb2p5/
step_002_Sb5/
step_003_Sb7p5/
...
In mode = "full", each step runs:
generate -> scan -> relax -> filter -> optional bandgap -> formation -> collect
After relaxation, the lowest-energy relaxed candidate is copied to:
step_xxx_<composition>/best_relaxed/POSCAR
This structure is then used as the base POSCAR for the next sequential step.
Modes
full
Runs the complete sequential workflow for each composition.
[sequential]
mode = "full"
recompute_energies
Reuses existing relaxed sequential structures and reruns only formation-energy evaluation and database collection.
This is useful when changing the thermodynamic reference, for example:
oxides_ref = ["Sb2O5"]
without regenerating, scanning, or relaxing structures again.
[sequential]
mode = "recompute_energies"
Outputs
Each step writes its own local database:
sequential_structures/step_001_Sb2p5/results_database.csv
sequential_structures/step_002_Sb5/results_database.csv
At the end, all step databases are merged into the project-root database:
results_database.csv
Bandgap Control
The bandgap stage can be skipped in sequential calculations using:
[bandgap]
enabled = false
When disabled, the final database is still written, but the bandgap_eV column
is left empty.