Model Database

The repository’s modeldb/ directory contains public CSV parameters for ion implantation, damage, and thermal annealing. The data are generic simulation- and literature-based parameters. Check the table headers and provenance before selecting them for a particular process.

Available data

Directory	Coverage in this checkout
implant/	Boron and phosphorus in crystalline silicon; dual-Pearson IV profiles at 8, 10, 30, 50, 100, 130, and 200 keV, with 7° tilt and 22° twist.
damage/	Boron and phosphorus damage profiles in silicon.
anneal/annealing.csv	Intrinsic diffusivities for B, P, Al, Ga, In, As, Sb, and Bi in silicon; additional solubility and effective damage-driven TED parameters for B and P.

Implant tables use nanometers, keV, degrees, and ions/cm². Anneal table diffusivity prefactors use nm²/s and activation energies use eV; solubility prefactors use cm⁻³. The C++ anneal lookup converts length-dependent parameters to the requested geometry unit.

The public tables interpolate within their supported ranges; they do not provide extrapolation to arbitrary energies or implant geometries. Additional anneal diffusivity rows do not imply matching implant or damage tables.

Configure the database location

Point to the directory containing implant, damage, and anneal:

#include <psModelDb.hpp>

viennaps::setModelDbRoot("/path/to/ViennaPS/modeldb");

import viennaps as vps

vps.setModelDbRoot("/path/to/ViennaPS/modeldb")

Alternatively, set the VIENNAPS_MODELDB_ROOT environment variable and call initModelDbRoot() (vps.initModelDbRoot() in Python). Initialization first uses the VIENNAPS_MODELDB_DIR compile-time definition if the root is empty, then applies a nonempty environment override. getModelDbRoot() returns the configured path. The C++ implantation example defines a source-tree default. For a separately installed application, configure a path to your own copy of the database rather than assuming a checkout is available.

C++ table-based implant setup

With a cell-set domain prepared as in the implantation guide:

#include <models/psImplantSetup.hpp>

viennaps::TableImplantRecipe<double> recipe;
recipe.species = "B";
recipe.material = "Si";
recipe.substrateType = "crystalline";
recipe.energyKeV = 10.0;
recipe.tiltDeg = 7.0;
recipe.rotationDeg = 22.0;
recipe.doseCm2 = 1e13;
recipe.screenThickness = 0.0;

auto setup = viennaps::makeTableImplant<double, 2>(recipe);
auto model = viennaps::SmartPointer<viennaps::IonImplantation<double, 2>>::New();
viennaps::applyImplantSetup(*model, setup);
viennaps::Process<double, 2>(domain, model, 0.0).apply();

The helper assigns species-specific field labels such as B_total, B_damage, and B_damage_last. dopantFields(species) also supplies matching active and defect field names for annealing.

The setup helpers shown above are C++ APIs. Python exposes analytical profiles, ImplantTableModel, DamageTableModel, and the IonImplantation/Anneal setters. For example, use an explicit table path to construct the dopant profile:

profile = vps.ImplantTableModel(
    vps.getModelDbRoot() + "/implant/boron_in_silicon_crystalline.csv",
    "B", "Si", "crystalline", 10.0, 7.0, 22.0, 1e13, 0.0)
implant = vps.IonImplantation()
implant.setImplantModel(profile)
implant.setDose(1e13)
implant.setTiltAngle(7.0)
implant.setLengthUnit(1e-7)

See the Python scripts in the implantation example for a complete table-based workflow with damage and annealing.

Custom data and lookup failures

Missing database paths, unsupported species/materials, and out-of-range conditions produce model-data errors. Use conditions covered by the tables, provide custom CSV files through the recipe’s tableFileName, or configure analytical profile moments and anneal parameters directly. Supplying an implant profile alone does not supply a damage model for a defect-coupled anneal.

The C++ helpers report lookup failures as viennaps::modeldb::ModelDbError. runWithModelDbErrors(...) is available for command-line programs that want a formatted diagnostic and exit code 2.

See the database files and anneal provenance for the parameter sources and table formats.