Whether using AMSET via the command-line or python API, the primary controls are contained in the settings file or dictionary. An example AMSET settings file is given here.
The settings are grouped into sections. The description for each section and settings parameter is given below.
All settings are also controllable via command-line flags. The corresponding command-line interface options are also detailed below. Any settings specified via the command-line will override those in the settings file.
These settings control the AMSET run, including interpolation density and temperature/doping ranges.
Controls which doping levels (carrier concentrations) to calculate. Concentrations given with the unit cm–3. Can be specified directly as a comma separated list, e.g.:
Alternatively, ranges can be specified using the syntax
For example, the same doping concentrations as above can be written::
Negative concentrations indicate electrons (n-type doping), positive concentrations indicate holes (p-type doping).
['1.e15', '1.e16', '1.e17', '1.e18', '1.e19', '1.e20', '-1.e15', '-1.e16', '-1.e17', '-1.e18', '-1.e19', '-1.e20']
Controls which temperatures to calculate. Temperatures given in Kelvin. Can be specified directly as a comma separated list, e.g.::
Alternatively, ranges can be specified using the syntax
For example, the same temperature range as above can be written::
Interpolation factor controlling the interpolation density. Larger numbers
indicate greater density. The number of k-points in the interpolated band
structure will be roughly equal to
interpolation_factor times the number
of k-points in the DFT calculation. This is the primary option for controlling
the accuracy of the calculated scattering rates. Transport properties should
be converged with respect to this parameter.
Which scattering mechanisms to calculate. If set to
auto, the scattering
mechanisms will automatically be determined based on the specified material
parameters. Alternatively, a comma separated list of scattering mechanism
can be specified. Options include:
ACD(acoustic deformation potential scattering)
IMP(ionized impurity scattering)
POP(polar optical phonon scattering)
CRT(constant relaxation time)
MFP(mean free path scattering)
ACD,IMP,POP. The scattering mechanism will only be calculated
if all the required material parameters for that mechanism are set. See the
scattering section <scattering>_ of the documentation for more details.
Path to wavefunction coefficients file. The coefficients can be extracted from a VASP WAVECAR using the command:
This command also requires the vasprun.xml to be in the same folder.
Use projections to calculate wavefunction overlap. This can often result in very poor performance, and so is not recommended.
In order to use projections, the VASP calculation must be performed with
LORBIT = 11.
The amount to scissor the band gap, in eV. Positive values indicate band gap opening, negative values indicate band gap narrowing. Has no effect for metallic systems.
Set the band gap to this value, in eV. Will automatically determine and apply the
correct band gap scissor for the specified band gap. Cannot be used in
combination with the
scissor option. Has no effect for metallic systems.
How to handle "zero-weighted" k-points if they are present in the calculation. Options are:
- keep: Keep zero-weighted k-points in the band structure.
- drop: Drop zero-weighted k-points, keeping only the weighted k-points.
- prefer: Drop weighted-kpoints if zero-weighted k-points are present in the calculation (useful for cheap hybrid calculations).
Whether free carriers will screen polar optical phonon and piezoelectric scattering rates. This modifies the matrix elements from a dependence to a dependence, where is the inverse screening length that depends on the temperature, carrier concentration, and high-frequency dielectric constant.
This can result in a large reduction in the scattering rates at high carrier concentrations.
These settings control the materials properties required to calculate the scattering rates.
The high-frequency dielectric constant, in units of . Can be given as a 3x3 tensor or a single isotropic value.
Required for: POP, PIE
The static dielectric constant, in units of . Can be given as a 3x3 tensor or a single isotropic value.
Required for: IMP, POP
The elastic constants as the full 3x3x3x3 tensor or 6x6 Voigt form, in GPa.
Alteratively, a single averaged value can be given (not recommended).
Required for: ACD, PIE
Path to file containing deformation potentials for all bands, generated
amset deform read.
Alternatively, Can be given as a comma separated list of two deformation potentials for the VBM and CBM, respectively in eV, e.g.:
Or a single value to use for all bands in metals.
Required for: ACD
The piezoelectric constants () in C/m2 given as either the full 3x3x3 tensor or the 3x6 Voigt form.
Required for: PIE
The charge of acceptor defects, in units of electron charge.
Required for: IMP
The charge of donor defects, in units of electron charge.
Required for: IMP
The polar optical phonon frequency, in THz. Generally, it is ok to take the highest optical phonon frequency at the Gamma point.
Required for: POP
Basic version of boundary scattering in which the scattering rate is set to , where is the group velocity and is the mean free path in nm.
Required for: MFP
A constant relaxation time to use as the minimum relaxation time for all k-points.
It is not recommended to use this option in conjunction with any other scattering rates. Instead, this should be used to compare against results calculated in the constant relaxation time approximation.
Required for: CRT
These settings control the speed and accuracy of calculated properties. In general the defaults should give converged values.
The energy cut-off used to determine which bands to include in the interpolation and scattering rate calculation, in eV.
The Fermi–Dirac derivative tolerance that controls which k-points to calculate the scattering for. Given as a percentage from 0 to 1. Larger values indicate that the fewer k-points will be calculated, smaller values indicate a larger portion of the Brillouin zone will be calculated.
The energy step for the calculated density of states and transport density of states, in eV. Controls the accuracy of determining the position of the Fermi level and transport properties. Smaller is better but can quickly get more expensive.
The symmetry finding tolerance, in Å.
Number of processors to use.
-1 indicates to use all available
When using multiprocessing it is recommended to run
export OMP_NUM_THREADS=1 before
Cache interpolated wavefunction coefficients. This means that the coefficients
for each band and k-point on the Fourier interpolated k-point mesh are only
calculated once. While this can yield a significant speed-up, it also massively
increases memory requirements, especially if using a low value of
fd_tol, or if
the system contains very flat bands.
If memory issues occur, it is recommended to set
These settings control the output files and logging.
Whether to calculate n- and p-type carrier mobilities. Has no effect for metallic systems where mobility is not well defined.
Whether to report the individual scattering rate mobilities. I.e., the mobility if only that scattering mechanism were present.
The output file format. Options are:
write_mesh=True is not supported using the
Whether to write the input settings to a file called
Whether to write the full k-dependent properties to disk. Properties include the band energy, velocity and scattering rate. Only k-points in the irreducible wedge are included.
Note: for large values of interpolation_factor his option can use a large amount of disk space.
Whether to print log messages.