Sentaurus Materials Workbench Reference Manual

Introduction

Sentaurus Materials Workbench helps you to create material references and to set up input files for atomistic calculations using DFT calculations, or empirical potentials, or both. It can automate defect generation and includes several techniques and recipes to increase the accuracy of calculations. Furthermore, from the atomistic calculation results, Sentaurus Materials Workbench analyzes the results and generates the Sentaurus model parameters that can be used for different TCAD tools, but in particular for dopant diffusion and reaction simulations with Sentaurus Process, and band structure generation needed for Sentaurus Device.

Note

The SMW scripts need to import the package explicitly. In other words, be sure to include a from SMW import * at the beginning of the scripts.

There are different modules or workflows under Sentaurus Materials Workbench. The following content describes its main capabilities. More technical descriptions and further information on SMW algorithms can be found at Sentaurus Materials Workbench Technical Descriptions.

Material specifications

Using MaterialSpecifications objects you can define computational settings for your calculations. Sentaurus Materials Workbench comes with a MaterialSpecificationsDatabase with predefined MaterialSpecifications for industry relevant materials. This database should be the starting point for most applications.

Bandstructure calibration

Class SentaurusBandstructureCalibration is used to obtain optimized parameters for bandstructure models (e.g., effective mass, k.p) available to the Sentaurus tools by calibrating model bands to first-principles values. The typical calibration workflow is:

  1. Atomistic nanowires or nanoslabs are built with SMW and its DFT bandstructure is calculated. Users input the desired crystal orientations and dimensions (height and width for nanowires, thickness for nanoslabs). All surfaces are passivated with hydrogen.
  2. Finite-element meshes, corresponding to the atomistic nanostructures, and model bandstructures are automatically calculated with Sentaurus tools. Sentaurus Process is used to create structure meshes and Sentaurus Band to calculate model bandstructures.
  3. SMW adjusts the model parameters (effective masses or k.p parameters) iteratively, following a conjugate-gradient minimization algorithm, until the band dispersion from Sentaurus Band matches the DFT bandstructure. Parameters for bulk silicon are used as initial guess.

Class SentaurusBandstructureCalibration supports Silicon rectangular nanowires (Wire) and nanoslabs (Slab). Calibration of both conduction (effective-mass model, SentaurusWireEffectiveMassModel) and valence bands (k.p model, SentaurusWireKdotPmodel) are available for nanowires. Only calibration of valence bands (SentaurusSlabKdotPmodel) is supported for supported.

Single defect specification and convergence studies

Formation energies and trap energy levels for a variety of defects and supercell sizes can be done with ChargedPointDefect. An overview of how to specify particular defects and perform convergence studies can be read at the ChargedPointDefect Notes.

The available defect classes are:

All to be run with:

Band diagram extraction

The class banddiagramextraction_c extract band diagram from first principle calculations of multilayer 2D structures. The averaged conduction/valence bandedge, bandgaps and work functions/electron affinity in each layer can be extracted. To extract work functions, vacuum energy needs to be calculated. In order to calculate vacuum energy for every layer, the multilayer structure are divided into single layers with vacuum. The outflow for band diagram extraction is outlined as follows:

  1. Built a multilayer 2D structure. Single layers are also created corresponding to each individual layers in the multilayer structure.
  2. Perform First principle calculation for multilayer and single layers. Banddiagram related properties such as effective potential, chemical potential, fat bandstructure, projected density of states are calculated.
  3. Extract vacuum energy, chemical potential, bandedges.
  • banddiagramextraction_c

Defect characterization and migration

A typical outflow for defect characterization and migration is outlined as follows:

  1. A reference system is created or used. Users set up DFT or force-field calculation input files for the reference host materials with no defect, perform convergence analysis, and find the proper setup that generates well converged results. Using this reference setup, users obtain the optimized lattice parameters and atomic positions. The ChargedPointDefect can be used to provide a structured and automated way to perform such convergence studies.
  2. Alternative, users can utilize a precalculated material reference already available in the Material specifications. This is the suggested start for faster, simpler calculations.
  3. Sentaurus Materials Workbench takes this reference setup and, as requested by the user, generates the calculation procedures for various possible defect structures in the supported crystal structures, running them as different tasks. It also generates the proper tasks for the lattice vibrational mode frequencies of the defect structures and transition states.
  4. Sentaurus Materials Workbench can also generate the migration paths between the above-obtained stable defect structures as instructed by users, and submits the path optimization nudged elastic band (NEB) tasks. It also checks the convergence and consistency of such calculations.
  5. Finally, Sentaurus Materials Workbench compiles the atomistic results and calculates the formation energies, migration energies, vibrational entropies and other related quantities. With such information it calculates the Sentaurus Process continuum and KMC model parameters from all of the DFT results.
  6. Sentaurus Materials Workbench compiles the atomistic results and calculates the formation energies. Sentaurus Materials Workbench also checks the convergence and consistency of the calculation.

Empirical potential calculation

Sentaurus Materials Workbench supports three different ways to automatically compute empirical potentials in compound materials. They are described in Chemical Potentials in Compound Materials.

Defect characterization

Defect characterization is performed using the different defect lists.

Examples and an overall view of the creation and use of defect lists can be read in Operating with defect lists.

For a detailed explanation on single defect characterization and/or convergence studies for different supercell sizes, please refer to ChargedPointDefect.

Information on already available materials and how to define the material specifications needed for different Sentaurus Material Workbench modules is available at MaterialSpecifications and Material specifications.

Specific information on how to create, characterize and perform defect calculations can be read at the following classes:

Although defects can be updated (run) one by one with the update member function, they can also be updated in one instructions using updateAllDefectsAndTransitions. This approach allows to run different defect types in parallel, see Notes.

Defect migration

Defect migration is performed by running Nudged Elastic Band (NEB) calculations between initial and final defec states. The calculations are automtically set up by the TransitionPathList class.

Details on the overall methodology are to be found at the transition path list Setting Up the Migration Paths.

Two modes of operation can be used.

  1. Regular NEB between initial and final defect states.
  2. Ring mechanism, only applicable for dopant-vacancy pairs (DefectPairList), where the migration of the vacancy around the dopant is also accounted, as explained in Dopant-Vacancy ring mechanism.

Paths can be updated one by one with the update member function, but can also be updated together with the defects using updateAllDefectsAndTransitions. See Notes.

Diffusion in amorphous materials

Sentaurus Materials Workbench can be used to estimate diffusion of impurities in amorphous materials. The procedure is described at Diffusivity calculation in amorphous materials.

Using defect characterization and migration values in Sentaurus Process

The calculations performed with the above lists can be used together with writeSentaurusParameters to produce Sentaurus Process input files. The extraction functions are:

Multilayer Builder

For an overall explanation on the multilayer builder features see Notes.

The available classes are:

GrainBoundaryScattering

The GrainBoundaryScattering calculates grain boundary resistances and specific resistivities of metal grain boundaries. For an overall explanation on the grain boundary scattering features see Notes.

The available classes are: