Surface Slabs and Adsorption Energies

Introduction

A clean infinite crystal does not expose a surface, but catalysts, electrodes and sensors act at surfaces. Periodic electronic-structure calculations represent a surface with a slab: several atomic layers repeated in two in-plane directions, with vacuum separating one slab from the next along the surface normal. The simple energy subtraction for adsorption is only meaningful when the slab, isolated adsorbate and combined system use compatible models. Layer thickness, vacuum, surface coverage, site choice and electrostatic image effects can all move the result by chemically important amounts.

Core explanation

For adsorption of one gas-phase molecule X on a clean surface S, a common electronic-energy definition is ΔEads = E(S+X) − E(S) − E(X). A negative value indicates that the combined slab-plus-adsorbate electronic energy is lower than the separated clean slab and isolated molecule under these references. This is not automatically an adsorption free energy at a specified temperature and pressure. The molecular gas has translational and rotational entropy that may be greatly reduced when bound; zero-point energy, thermal corrections, pressure, solvent and surface vibrations may matter. If X dissociates, the balanced chemical reference must be changed accordingly. An adsorption energy for oxygen atoms referenced to O₂ gas, for example, uses the appropriate stoichiometric fraction of E(O₂), with known method sensitivity for that reference taken into account.

Slab thickness determines whether the middle layers resemble bulk material. A slab that is too thin can couple its top and bottom surfaces or distort the electronic states near adsorption. Vacuum spacing reduces interaction between repeated slabs along the surface normal; insufficient vacuum creates artificial electrostatic or density overlap. The in-plane cell sets adsorbate coverage and separation between its periodic images. One molecule in a small cell describes a high-coverage ordered overlayer, not an isolated adsorbate. For polar or asymmetric slabs, repeated dipoles can make the energy converge slowly with vacuum size. VASP's developer guidance on slab electrostatic corrections describes the need for vacuum and dipole treatment when appropriate.

The clean and adsorbed slab calculations should use the same cell vectors, layer count, surface termination, core potentials, functional, cutoff, k-point scheme and constraints. The isolated molecular reference may require a separate large box and an appropriate spin state. At fixed slab geometry, ΔEads isolates bond formation and electronic rearrangement plus any adsorbate deformation. If the surface is separately relaxed for clean and adsorbed states, the difference also includes surface reconstruction and strain relaxation; both definitions can be useful but should not be mixed. A practical surface workflow tests multiple adsorption sites and orientations because a geometry optimizer finds a local minimum near its starting structure, not necessarily the global preferred site.

Charge and solvent introduce further complexity. Charged slabs in a conventional periodic cell can have severe long-range electrostatic artifacts, and an electrochemical potential cannot be inferred from one raw charged-cell total energy. An aqueous adsorbate competes with water and ions for the surface; a vacuum binding energy may not rank coverages in an electrolyte. A thermodynamic surface model may use chemical potentials and potential-dependent free energies to compare phases. A surface adsorption tutorial from the VASP developers illustrates the clean slab, adsorbate and combined-system setup with vacuum and dipole controls. The broader lesson transfers to other periodic codes, though exact settings differ.

Basis-set and electronic-method limits remain. A plane-wave cutoff and k mesh must be converged for the adsorption difference , especially if site preferences are small. Dispersion may matter for physisorption, while strongly correlated or open-shell adsorbates can stress a semilocal density functional. The adsorbate's charge and spin state may change on binding. Report a well-defined sign convention: some communities call the positive desorption energy “binding energy,” whereas others call the negative formation difference “adsorption energy.” The formula beside the number removes ambiguity.

Step-by-step reasoning

1. Select the crystallographic face, termination and slab thickness for the real surface question. 2. Add sufficient vacuum and choose an in-plane cell corresponding to the intended coverage. 3. Converge clean-slab and adsorbed energies with respect to layers, vacuum, lateral cell, cutoff and k points. 4. Optimize several plausible adsorption sites while keeping reference slab settings consistent. 5. Compute ΔEads from the explicit balanced formula and state whether surface relaxation is included. 6. Add zero-point, entropy, pressure, solvent or potential-dependent terms only for the target free-energy question.

Visual explanation

Draw three horizontal atomic layers in a rectangular cell with an X molecule above the top layer. Show neighboring copies to the left and right and a second slab above, separated by vacuum. Mark the lateral cell length as controlling coverage and the vertical gap as vacuum spacing. Beneath, write three energy boxes—combined slab, clean slab, gas X—with arrows indicating E(S+X) − E(S) − E(X). A dipole arrow normal to an asymmetric slab indicates where image electrostatics can arise.

Real-world analogy

Testing how a guest sits on a sofa depends on the sofa's width, stiffness and whether another copy of the guest is squeezed beside them. A surface supercell similarly fixes available area and repeated neighbors. Adding more empty space above the sofa does not give the guest more room sideways; vacuum and lateral-cell convergence solve different artifacts. The analogy is geometric, not a model of electrons or chemical bonds.

Real-world example

A catalyst researcher studies CO on a nickel surface. Calculations on a small cell predict a preferred site, but increasing lateral cell size changes the adsorption energy because periodic CO molecules interact and coverage falls. A thicker slab changes the site difference slightly, while adding a dipole correction stabilizes the asymmetric slab energy with vacuum width. The researcher reports the surface face, coverage, spin treatment and energy formula. To compare with temperature-programmed desorption or reaction conditions, the researcher also estimates gas entropy and competing surface states rather than equating ΔEads to an observed equilibrium constant.

Why?

Why is vacuum spacing not enough to model an isolated adsorbate? Periodic repetition occurs in all directions. Vacuum separates slab images vertically, but the molecule is still copied across the surface in the two in-plane directions. If the cell is small, neighboring adsorbates interact laterally and the computed state represents a dense ordered layer. To approach dilute coverage, enlarge the in-plane area and check the adsorption energy's convergence.

Common misconception

“Negative adsorption energy proves the surface will be covered under all conditions.” Entropy and chemical potentials determine coverage. “Any large-looking vacuum is converged” ignores long-range dipoles and charge. “One optimized site establishes the preferred geometry” overlooks other local minima and surface reconstructions. “Adsorption energy and desorption barrier are the same quantity” is false when the pathway has an additional transition state or intermediate.

Worked example

Suppose hypothetical total energies are E(S+X) = −105.80 eV, E(S) = −100.00 eV and E(X) = −5.20 eV, computed with compatible references. Then ΔEads = −105.80 − (−100.00) − (−5.20) = −0.60 eV. With 1 eV ≈ 96.49 kJ mol⁻¹, this is about −58 kJ mol⁻¹ per adsorbed molecule. If increasing the lateral cell changes it to −0.48 eV, the 0.12-eV difference signals coverage or image effects and should be understood before reporting an isolated-site value. Neither number includes the gas molecule's entropy at a specified pressure or a solvent environment.

Quick check

1. In ΔEads = E(S+X) − E(S) − E(X), what does a negative result mean? Answer: The combined surface-plus-X electronic state is lower in energy than the stated separated references. 2. Which cell dimension primarily controls the artificial lateral separation of repeated adsorbates? Answer: The in-plane surface cell dimensions, not the vertical vacuum width.

Exam focus

Write the adsorption-energy formula before calculating a sign. Identify layer thickness, vacuum and lateral area as separate convergence variables. State what a clean slab and an isolated molecule reference must share with the adsorbed system. Distinguish electronic adsorption energy from a Gibbs free energy and from a kinetic desorption barrier. Explain why asymmetric or polar slabs may need electrostatic correction and convergence checks.

Advanced insight

Surface phase diagrams compare structures as functions of adsorbate chemical potential, which depends on gas pressure or solution conditions; the lowest energy at one fixed stoichiometry is not the whole operating-phase prediction. Adsorbate-induced reconstruction may make a separately relaxed clean slab the appropriate reference for equilibrium, while a frozen-slab comparison can isolate local bond formation. Charged interfaces may need explicit counterions, constant-potential approaches or specialized finite-size corrections. Error cancellation in ΔEads can be good when all slab calculations are closely matched, but gas-phase reference errors and functional-dependent dispersion can remain large.

Summary

A periodic slab models a surface through repeated atomic layers and vacuum. Adsorption energies compare a combined slab with a clean slab and balanced adsorbate reference under one clear sign convention. Layer, vacuum, lateral coverage, site, cutoff and k-point convergence all matter. Polar images and electronic-method limits can change small energy differences. To predict real coverage or catalysis, extend the electronic energy to the relevant thermodynamic and kinetic conditions.

Practice questions

1. If E(S+X) = −62.4 eV, E(S) = −59.0 eV and E(X) = −3.0 eV, what is ΔEads? Answer: −0.4 eV using the stated formula. 2. Why does an asymmetric slab often converge slowly with vacuum width? Answer: Its repeated dipoles interact electrostatically across the periodic vacuum. 3. Does a thicker vacuum reduce adsorbate–adsorbate interaction across the surface? Answer: Not the lateral interaction; a larger in-plane cell is needed for that. 4. Why might an exothermic adsorption electronic energy correspond to weak coverage at high temperature? Answer: Adsorption loses gas translational entropy, so the Gibbs free energy and equilibrium coverage depend on temperature and pressure.