Example: Cu_slab: 2-dim. Finite temperature and orbital plot¶

Download Cu_slab.run

A two-dimensional infinite (periodic boundary conditions) slab calculation is performed for Cu. The dimensionality is simply defined by the number of records in the Lattice data block. In a 2-dimensional calculation the lattice vectors are put in the xy-plane. (In a one-dimensional calculation (polymer), the lattice vector is taken along the x-axis in the program.)

Slab calculations for metals frequently suffer from SCF convergence problems, as a result of the open valence band(s). To help the program converge it is often useful or even necessary to use some special features, such as the ElectronicTemperature key. This particular key requires a numerical value (0.025 in the example) and implies that a finite-temperature electronic distribution is used, rather than a sharp cut-off at the Fermi level. The numerical value is the applied energy width, in hartree units.

The so-modified electronic distribution also affects the energy. The ‘true’ zero-T energy is computed, approximately, by an interpolation formula. The interpolation is not very accurate and one should try to use as small as possible values for the ElectronicTemperature key so as to avoid increasing uncertainty in the results. The program prints, in the energy section of the output file, the finite-T correction term that has been applied through the interpolation formula. This gives at least an indication of any remaining uncorrected deviation of the outcome from a true zero-T calculation.

In the second run the runkf file of the first run is used to do an orbital plot restart. Normally you would rename the resulting TAPE41 to “myslab.t41” and watch the orbitals with adfview.

# ----------------------------- first run --------------------------

$ADFBIN/band << eor
Title Cu slab

Comment
 Technical
   Quadratic K space integration
   Good real space integration accuracy
 Features
   Lattice   : 2D
   Unit cell : 1 atom, 1x1
   Basis     : NO+STO w/ core
   Options   : ElectronicTemperature (temperature effect)
End

Kspace 5
Accuracy 4

Convergence
   ElectronicTemperature  0.025
End

Lattice
   4.822 0.0
   0.0   4.822
End

Atoms
   Cu  0.0   0.0   0.0
End

BasisDefaults
   BasisType DZ
End

EndInput
eor

mv RUNKF CuSlab.runkf
rm Points

# ----------------------------- orbital plot --------------------------
export NSCM=1

$ADFBIN/band -n 1 << eor
Title Cu slab orbital plot

Comment
 Technical
   Quadratic K space integration
   Good real space integration accuracy
 Features
   Lattice   : 2D
   Unit cell : 1 atom, 1x1
   Basis     : NO+STO w/ core
   Options   : ElectronicTemperature (temperature effect)
End

Kspace 5
Accuracy 4

Restart
   File CuSlab.runkf
   OrbitalPlot
End

Grid
   Type Coarse
End

OrbitalPlot
   1 Band 2 4 ! k-point 1, bands 2 to 4
   3  -0.1 0.1 ! k-point 3 orbitals within 0.1 Hartree from Fermi Level
End

Convergence
   ElectronicTemperature  0.025
End

Lattice
   4.822  0.000
   0.000  4.822
End

Atoms
   Cu  0.0   0.0   0.0
End

BasisDefaults
   BasisType DZ
End

EndInput
eor

echo "Begin TOC of tape41"

$ADFBIN/dmpkf -n 1 TAPE41 --toc

echo "End TOC of tape41"