Skip to content

3.4 Slabs, adsorption, dipoles and work functions

These are unexecuted teaching inputs and starting settings to test. Original figures are schematics, not computed results. Use licensed VASP and PAW data, replace every placeholder, record the executable version and validate convergence.

3.4.1 Model, units and provenance

Use eV for energy, Å for length and eV/Å for force; 1 kbar = 0.1 GPa. State normalization per atom, molecule, primitive cell or simulation cell. Record PAW identifiers, release, ZVAL, ENMAX and permitted hashes; never redistribute POTCAR. SCF convergence addresses the chosen electronic problem; convergence of the target property requires separate tests.

 Slabs, adsorption, dipoles and work functions — original conceptual schematic

Original schematic. Curves explain concepts; blank data areas await verified learner results. No calculation is claimed.

\[ \Phi=V_{\mathrm{vac}}-E_F,\qquad\gamma=\frac{E_{\mathrm{slab}}-NE_{\mathrm{bulk/atom}}}{2A}\quad\text{(equivalent clean faces)} \]

3.4.2 Worked case: procedure, interpretation and checks

Goal and intuition

A surface calculation retains periodicity parallel to a surface and inserts vacuum normal to it. The slab has two sides, finite thickness and periodically repeated adsorbates. Thus “one molecule on a surface” actually means a particular coverage and repeated surface pattern. A work function additionally needs a field-free vacuum reference; the Fermi energy printed for a bulk cell alone is insufficient.

Use the validated Al bulk lattice to build an Al(111) slab for a first clean-surface study, then add one neutral H₂O as an adsorption bookkeeping exercise. This is a methodological example, not a claim that a chosen water geometry is the stable adsorption structure. Prerequisites: bulk relaxation, metallic sampling, isolated-molecule references and visualization of periodic cells. A reputable structure builder may construct the slab, but inspect atom count, termination, orientation and nearest-neighbor distances yourself.

Workflow

  1. Build a centered neutral slab with a clearly defined surface normal along the third lattice vector. Start, for example, from a 2×2 lateral cell and several atomic layers, but treat thickness and area as independent convergence variables.

Unexecuted slab settings

The Γ-only normal-direction sampling is a model-specific starting choice for an adequately separated slab, not a substitute for testing the vacuum. Adding O/H may raise the largest ENMAX. Use the same explicit chosen cutoff across all members of the adsorption-energy comparison rather than leaving each run to choose a different default. Dipole corrections do not remove lateral adsorbate interactions or eliminate the need for vacuum convergence.

Energy bookkeeping

Define Eads = E(slab+molecule) − E(clean slab) − E(isolated molecule), with negative values indicating favorable adsorption under this electronic-energy convention. Keep slab atom count, lateral cell, PAW data, functional and numerical accuracy compatible. Say whether the clean slab is separately relaxed under the same constraints and whether the molecule reference is its relaxed gas-phase geometry. If you instead use fragments frozen in the adsorbed geometry, you calculate a different decomposition involving deformation energies.

For a symmetric clean slab with two equivalent surfaces, a common surface-energy expression is γ = [Eslab − N Ebulk/atom]/(2A). The factor of two counts the two equivalent surfaces. Do not apply it blindly to asymmetric or nonstoichiometric slabs. For polar, charged, multicomponent or reconstructed surfaces, the thermodynamic references and electrostatics require a more advanced treatment.

Work-function interpretation

Plane-average the selected electrostatic potential over the surface plane, obtaining V(z). Find a genuinely flat, charge-free vacuum region and read its plateau Vvac. For a metallic surface, Φ = Vvac − EF, using energies in consistent units and the Fermi level from the same calculation. An asymmetric slab can have different plateaus and work functions on its two sides; report which side corresponds to adsorption. The correction discontinuity is not a physical vacuum plateau and should not be averaged into one.

The current VASP work-function guide recommends the Hartree-plus-ionic potential, obtainable with LVHAR, rather than relying on a slowly decaying exchange-correlation contribution in vacuum. The optional automated vacuum-level extraction should be checked against the profile rather than trusted blindly. For semiconductors, Fermi-level placement, doping and band bending make interpretation more subtle than the simple metallic exercise.

Convergence experiment and limitations

Vary slab thickness, total vacuum gap, lateral area and in-plane k density independently. Track adsorption energy, adsorption height, middle-layer geometry and each side's work function. A single vacuum plateau does not prove slab thickness is sufficient. A converged adsorption energy in a 2×2 cell is a result at that periodic coverage, not automatically the isolated-adsorbate limit. A water adsorption energy from semilocal DFT alone may miss relevant dispersion physics; changing the dispersion model is a physical-method comparison and must be applied consistently to all references.

Exercise and original figure

Submit a surface-model audit showing Miller index, lattice source, slab layers, fixed layers, area, vacuum gap, adsorbate coverage, charge, dipole settings and energy reference cycle. Include a V(z) plot with the chosen plateau interval and EF labeled, and explain any inequivalence between the two sides. If no reliable plateau exists, the valid deliverable is a diagnosed failure and next test, not a work-function number.

3.4.3 Unexecuted inputs and analysis scaffolds

These are unexecuted teaching inputs and starting settings to test. Original figures are schematics, not computed results. Use licensed VASP and PAW data, replace every placeholder, record the executable version and validate convergence.

3.4.3.1 Input block 1

SYSTEM = Neutral asymmetric slab study
ENCUT = <ECUT_FOR_ALL_SPECIES_EV>
PREC = Accurate
EDIFF = 1E-7
LREAL = .FALSE.
ISMEAR = 1
SIGMA = 0.10
IBRION = 2
NSW = 150
ISIF = 2
EDIFFG = -0.02
LDIPOL = .TRUE.
IDIPOL = 3
DIPOL = 0.5 0.5 0.5
# DIPOL assumes a deliberately centered slab; inspect the actual geometry
# Use selective-dynamics flags in POSCAR for the stated layer constraints

3.4.3.2 Input block 2

Slab starter k mesh, must be converged in-plane
0
Gamma
6 6 1
0 0 0

3.4.3.3 Input block 3

# Separate final static electrostatic-potential stage
IBRION = -1
NSW = 0
LVHAR = .TRUE.
# Keep the appropriate dipole correction and converged settings
# Optional LVACPOTAV = .TRUE. documented for versions after VASP 6.4.3: verify installed version

3.4.5 Technical sources