Recent Posts71
General Questions and Answers / Re: Question about Defect Diffusion Workflow Template
« Last post by Vaida Arcisauskaite on September 16, 2025, 10:23 »Hello,
Thank you for sharing your input and output files for the defect diffusion calculations. Upon review, we noticed that the workflow does not include the calculation of the chemical potential for the dopant atom (Boron). As a result, while the calculation completes, the relevant properties cannot be extracted.
To resolve this, we recommend the following steps (see the attached pdf for instructions with screenshots from the Workflow Builder):
1. Add a ‘Block of Blocks’ for the Boron Reference Material.
2. Include the calculated B chemical potential in the ChemicalPotentialTable.
3. Left-click on the ChemicalPotentialTable block to specify the calculator and configurations for the chemical potential calculation.
4. If you wish to include vibrational corrections, specify the phonon_calculator (your trained MachineLearnedForce Field MTP). We recommend using MTP for vibrational corrections instead of DFT to make the calculation feasible.
Additionally, we suggest reviewing all block connections in the workflow by left-clicking on each block to verify and adjust settings as needed.
Other observations:
1. Isotropic Finite Size Corrections
• If you calculate Optical Spectrum and Elastic constants to extract dielectric constant and bulk modulus values, respectively, that would then be used as parameters in the calculation of Isotropic FiniteSize Corrections, then you need to specify this in the IsotropicFiniteSizeCorrection parameters block.
• We also recommend specifying a “higher” k-point sampling density for Optical Spectrum.
2. Pristine Material
• Set the “opgeom” under the PristineMaterial to Conventional Unit Cell (8 atoms).
• Then in the PristineConfiguration block specify Supercell Repetitions 64-atom (2×2×2) or 216-atom (3×3×3) supercells.
3. CPD simulation settings
• Use tighter DFT-LCAO SCF convergence criteria (e.g., 10⁻⁵). Increase the number of max iterations to 200.
• Increase the density mesh cutoff in DFT-LCAO — check convergence of total energy with respect to this parameter for the conventional unit cell.
• Tighten the maximum force tolerance for all geometry optimizations to 0.005.
Please let us know if you have any questions or need further assistance with implementing these changes.
Best regards,
Vaida
Thank you for sharing your input and output files for the defect diffusion calculations. Upon review, we noticed that the workflow does not include the calculation of the chemical potential for the dopant atom (Boron). As a result, while the calculation completes, the relevant properties cannot be extracted.
To resolve this, we recommend the following steps (see the attached pdf for instructions with screenshots from the Workflow Builder):
1. Add a ‘Block of Blocks’ for the Boron Reference Material.
2. Include the calculated B chemical potential in the ChemicalPotentialTable.
3. Left-click on the ChemicalPotentialTable block to specify the calculator and configurations for the chemical potential calculation.
4. If you wish to include vibrational corrections, specify the phonon_calculator (your trained MachineLearnedForce Field MTP). We recommend using MTP for vibrational corrections instead of DFT to make the calculation feasible.
Additionally, we suggest reviewing all block connections in the workflow by left-clicking on each block to verify and adjust settings as needed.
Other observations:
1. Isotropic Finite Size Corrections
• If you calculate Optical Spectrum and Elastic constants to extract dielectric constant and bulk modulus values, respectively, that would then be used as parameters in the calculation of Isotropic FiniteSize Corrections, then you need to specify this in the IsotropicFiniteSizeCorrection parameters block.
• We also recommend specifying a “higher” k-point sampling density for Optical Spectrum.
2. Pristine Material
• Set the “opgeom” under the PristineMaterial to Conventional Unit Cell (8 atoms).
• Then in the PristineConfiguration block specify Supercell Repetitions 64-atom (2×2×2) or 216-atom (3×3×3) supercells.
3. CPD simulation settings
• Use tighter DFT-LCAO SCF convergence criteria (e.g., 10⁻⁵). Increase the number of max iterations to 200.
• Increase the density mesh cutoff in DFT-LCAO — check convergence of total energy with respect to this parameter for the conventional unit cell.
• Tighten the maximum force tolerance for all geometry optimizations to 0.005.
Please let us know if you have any questions or need further assistance with implementing these changes.
Best regards,
Vaida