Defect Formation Energy and Equilibrium Concentrations in Oxides: From Thermodynamic Energetics to Defect Populations
Defect Formation Energy and Equilibrium Concentrations in Oxides: From Thermodynamic Energetics to Defect Populations
Point defects are central to the properties of functional oxides. Oxygen vacancies can support ionic transport, electronic defects can produce n-type or p-type conductivity, and aliovalent dopants can change the concentration of compensating defects. A quantitative description requires more than identifying which defects are chemically possible: it requires determining how energetically favourable each defect is and what concentration is established at equilibrium.
Defect formation energy provides an energetic measure for creating a defect under specified chemical and electronic conditions. At finite temperature, the equilibrium population is governed by formation free energy and configurational statistics, together with charge neutrality, site balance, and the chemical potentials imposed by the environment.
This article develops the connection between defect energetics and equilibrium concentrations. It distinguishes zero-temperature or static formation energies from finite-temperature formation free energies, explains the role of chemical potentials and charge states, and shows why a Boltzmann expression alone is not a complete defect-chemistry model.
1. What is a point-defect formation energy?
A point defect is a local departure from the ideal periodic crystal. Examples include vacancies, interstitials, substitutional dopants, antisite defects, and electronic carriers.
The formation energy is the energy cost associated with creating a defect relative to a chosen reference crystal and reservoirs of atoms and electrons. For a defect (D) in charge state (q), a widely used expression from atomistic calculations is
[ E_f(D^q)= E_{}(D^q)-E_{}() -i n_ii +q(E_F+E{}+V) +E{} ]
Here:
- (E_{}(D^q)) is the total energy of the defective supercell.
- (E_{}()) is the total energy of the reference bulk supercell.
- (n_i) is the number of atoms of element (i) added to the supercell; removal corresponds to a negative value under this convention.
- (_i) is the chemical potential of the corresponding atomic reservoir.
- (q) is the defect charge state, using the convention in which positive (q) means electrons have been removed.
- (E_F) is the Fermi level referenced to the valence-band maximum.
- (V) is a potential-alignment correction, when required.
- (E_{}) represents finite-size electrostatic and other correction terms appropriate to the calculation.
The exact bookkeeping convention must be checked in any computational study. Different publications may define atom-count changes or reference energies differently. The physical result is independent of notation when all terms are treated consistently.
For a neutral defect, (q=0), so the explicit Fermi-level term vanishes. Charged defects generally have formation energies that depend on the Fermi level.
2. Formation energy is conditional, not an intrinsic universal number
A defect formation energy is not generally a single fixed property of a material. It depends on the thermodynamic conditions and the reference choices used to define it.
Important dependencies include:
- Chemical potentials: oxygen-rich and oxygen-poor conditions change the energetic cost of creating oxygen-related defects.
- Fermi level: charged defects become more or less favourable as the electron chemical potential changes.
- Charge state: the same structural defect may exist in more than one charge state.
- Crystal structure and composition: local bonding, strain, dopant arrangement, and phase identity affect the energy.
- Computational methodology: supercell size, exchange-correlation functional, band-gap error, electrostatic corrections, and structural relaxation affect calculated values.
A reported formation energy should therefore be accompanied by its chemical-potential limits, Fermi-level reference, charge state, and computational convention.
3. From formation energy to equilibrium concentration
Consider a dilute population of independent defects that can occupy (N) equivalent sites. Suppose each defect has a formation free energy (G_f), and let (c) be the fraction of available sites occupied by the defect.
The number of ways to arrange (Nc) defects among (N) sites is
[ W= ]
The configurational entropy is
[ S_{}=k_BW ]
Using Stirling’s approximation for large (N),
[ S_{} -Nk_B]
If defects are independent and have the same formation free energy, the free-energy contribution is
[ G(c)=NcG_f-TS_{} ]
At equilibrium, minimize (G) with respect to (c):
[ =0 ]
which gives
[ G_f+k_BT()=0 ]
Therefore,
[ ]
For a dilute defect population, (c), so (1-c), yielding
[ ]
and the concentration is approximately
[ [D]N (-) ]
This is the familiar Boltzmann form. It is a useful starting point, but it assumes equivalent sites, dilute defects, independent defect formation, and a well-defined formation free energy.
If there are (g) equivalent configurations or orientations per site, the degeneracy contributes a statistical factor. Depending on how the formation free energy is defined, this can be written as a prefactor (g) multiplying the dilute concentration. Degeneracy must not be counted twice if it is already included in the free energy.
4. Why formation free energy is more general than formation energy
A static formation energy is often calculated from relaxed total energies, usually at or near zero temperature. An equilibrium concentration at finite temperature is governed by a formation free energy:
[ G_f(T)=H_f(T)-TS_f(T) ]
The enthalpy term includes energetic contributions such as local structural relaxation and, where included, vibrational contributions. The entropy term can include vibrational, configurational, electronic, and magnetic contributions, depending on the material and model.
The dilute equilibrium concentration becomes
[ [D]N ]
A frequently used approximation is to substitute a static formation energy for (G_f). This can be reasonable when the omitted entropy contributions are small or when the calculation is intended as a first estimate. It should not be interpreted as a complete finite-temperature treatment.
A moderate entropy contribution can substantially change the concentration at high temperature because it appears divided by (k_BT) in the exponential.
5. Chemical potentials and oxygen-rich or oxygen-poor conditions
The atomic chemical potentials in a defect formation expression describe exchange with external reservoirs. In an oxide, the oxygen chemical potential is especially important.
For an oxygen vacancy, an oxygen atom is removed from the crystal. Under the convention in the formation-energy expression above, (n_O=-1), so the oxygen-reservoir contribution is
[ -n_O_O=+_O ]
Thus, increasing the oxygen chemical potential raises the formation energy of an oxygen vacancy, all else being equal. Lower oxygen chemical potential makes oxygen removal more favourable.
For an ideal oxygen gas in equilibrium with the solid,
[ O(T,p{}) = _{}^(T) +k_BT () ]
Here, (_O) is the chemical potential per oxygen atom, and (p^) is the standard pressure. The expression assumes ideal-gas behaviour; fugacity should replace pressure when non-ideality matters.
As oxygen partial pressure decreases, (_O) decreases. The formation energy of an oxygen vacancy therefore tends to decrease, favouring reduction. This is the thermodynamic origin of the strong dependence of oxygen-defect populations on atmosphere.
Chemical potentials cannot be varied arbitrarily. They are constrained by phase stability: the selected values must not make the host oxide unstable against decomposition or precipitation of competing phases.
6. Charged defects and Fermi-level dependence
For a charged defect, the formation energy contains a term proportional to the Fermi level:
[ E_f(Dq)=E_f(Dq;E_F=0)+qE_F ]
where the reference is the valence-band maximum and other correction terms are understood to be included consistently.
Consequently:
- Positively charged defects become less favourable as (E_F) rises.
- Negatively charged defects become more favourable as (E_F) rises.
- Neutral defects have no explicit linear Fermi-level dependence.
The stable charge state at a given Fermi level is generally the one with the lowest formation free energy, after all relevant corrections are applied. A change in the lowest-energy charge state can occur at a thermodynamic charge-transition level.
This creates a self-consistency problem. The Fermi level controls charged-defect formation energies, while the concentrations of charged defects and electronic carriers determine the Fermi level through charge neutrality.
A defect calculation that assumes a Fermi level without checking charge neutrality may therefore predict concentrations that are not self-consistent with the material’s composition.
7. Charge neutrality determines the Fermi level
For a bulk material, the net charge must vanish on a macroscopic scale. In a simplified oxide containing doubly positively charged oxygen vacancies, electrons, holes, and singly negatively charged acceptors,
[ 2[V_O^{}]+[h^] = [n]+[A’] ]
This equation must be expanded to include all relevant charged defects and dopants.
For each charged defect, its concentration depends on its formation free energy and therefore, in general, on (E_F). The electronic carrier concentrations also depend on (E_F) and temperature.
The equilibrium solution is obtained by solving the defect-concentration relations together with charge neutrality and any required site-balance and composition constraints.
This is why a single isolated Boltzmann expression is not enough to determine the equilibrium defect population in a real oxide.
8. Example: a neutral vacancy population and the role of energy
Consider an illustrative dilute defect model with one type of neutral defect, one equivalent site per formula unit, and a formation free energy of (1.00 ). At (T=1000 K),
[ k_BT=(8.617{-5} \mathrm{eV,K{-1}})(1000 K) =0.08617 ]
The dilute site fraction is
[ c (-) ]
[ ]
If the formation free energy were instead (0.80 ) at the same temperature,
[ c (-) ]
[ ]
A reduction of only (0.20 ) increases the predicted dilute concentration by roughly an order of magnitude.
This example illustrates the exponential sensitivity of defect concentrations to formation free energy. It is not a prediction for a particular oxide: real charged defects require chemical-potential and Fermi-level terms, and the total population must satisfy charge neutrality.
9. Multiple defect species and competing populations
A material may contain several defects on different sublattices. Each species has its own formation free energy and available site population.
In the dilute limit, a simple expression for defect species (i) is
[ c_ig_i (-) ]
where (g_i) is an appropriate degeneracy or configurational prefactor, provided it is not already included in (G_{f,i}).
The total defect population is not obtained by selecting only the lowest formation energy without considering site availability, charge state, chemical composition, and charge neutrality. A defect with a somewhat higher formation energy may become important if it has a large site population or if its charge state is favoured by the self-consistent Fermi level.
At high concentrations, defects may interact, associate, order, or alter the local structure. The independent-defect approximation then becomes inadequate.
10. Dopants, compensation, and defect association
Aliovalent dopants change the thermodynamic balance of charged defects. For example, a trivalent dopant on a tetravalent cation site in a simple oxide is represented as a singly negative acceptor:
[ ]
Charge compensation may involve positively charged oxygen vacancies:
[ 2+ ]
This is a charge-neutral combination. If oxygen vacancies provide the dominant compensation mechanism, the nominal vacancy concentration can be constrained by the acceptor concentration.
However, the total vacancy concentration is not necessarily the concentration of mobile, unassociated vacancies. Dopant–vacancy association can produce neutral or partially compensated defect complexes. Such association changes the free-defect population and can affect ionic conductivity even when the total oxygen deficiency changes little.
A rigorous model may need separate formation free energies for isolated defects and defect complexes, together with association equilibria.
11. Equilibrium concentration is not the same as defect mobility
Formation energetics determine how many defects are thermodynamically present at equilibrium. Migration energetics determine how readily those defects move.
For a vacancy-mediated ionic conductor, the conductivity depends on both the concentration of mobile charge carriers and their mobility. In a simplified description,
[ =nq ]
where (n) is the mobile carrier concentration, (q) is the carrier charge, and () is the mobility.
For ionic defects, the effective carrier concentration may be lower than the total defect concentration if some defects are trapped in complexes, ordered arrangements, or immobile regions. Mobility depends on migration barriers, attempt frequencies, correlations, and microstructure.
A low vacancy formation energy does not automatically imply high ionic conductivity. A material may contain many vacancies but exhibit poor transport if the migration barrier is high or vacancy association is strong.
12. Computational and experimental approaches
First-principles calculations
Density-functional-theory calculations can estimate defect formation energies for specified structures, chemical potentials, charge states, and Fermi levels. A reliable workflow should consider:
- Supercell-size convergence.
- Structural relaxation around the defect.
- Electrostatic finite-size corrections for charged defects.
- Potential alignment and reference-energy consistency.
- Band-gap errors and the placement of band edges.
- Competing charge states and thermodynamic transition levels.
- Chemical-potential limits imposed by phase stability.
- Vibrational contributions when finite-temperature accuracy is required.
The resulting formation-energy diagrams can be used as inputs to a self-consistent defect-chemistry model.
Thermogravimetry and oxygen nonstoichiometry
Thermogravimetric measurements under controlled oxygen partial pressure can provide oxygen nonstoichiometry as a function of temperature and atmosphere. These data constrain defect models, but the measured mass change does not uniquely identify every defect species.
Combining thermogravimetry with conductivity, coulometric titration, oxygen-partial-pressure dependence, or spectroscopic measurements can improve the identification of dominant compensation mechanisms.
Electrical measurements
Electronic and ionic conductivity measurements can help constrain carrier populations and transport mechanisms. However, conductivity alone generally cannot determine a unique defect concentration because mobility, grain boundaries, secondary phases, and defect association also contribute.
13. A practical workflow for predicting equilibrium defect concentrations
- Define the host phase and composition. Identify the crystallographic sites and possible defect species.
- Specify external conditions. Set temperature, oxygen partial pressure, water activity, and other relevant chemical reservoirs.
- Establish chemical-potential limits. Ensure that the chosen reservoirs are compatible with host-phase stability.
- Calculate or obtain formation free energies. Include charge states and Fermi-level dependence where relevant.
- Choose a statistical model. Decide whether dilute independent defects are adequate or whether interactions and site exclusion must be included.
- Write mass-action and site-balance relations. Use consistent activities or normalized site fractions.
- Impose charge neutrality. Solve for the self-consistent Fermi level and charged-defect populations.
- Include association or ordering if necessary. Distinguish total defects from free, mobile defects.
- Compare with experiments. Test predicted oxygen nonstoichiometry, conductivity trends, and atmosphere dependence.
- State the model limits. Report assumptions, standard states, computational corrections, and uncertainty.
14. Common mistakes
- Treating formation energy as a universal material constant. It depends on chemical potentials, charge state, Fermi level, and reference conventions.
- Using a zero-temperature energy as an exact finite-temperature free energy. Vibrational and other entropy contributions may matter.
- Ignoring the Fermi-level dependence of charged defects. This can produce non-self-consistent concentrations.
- Applying the Boltzmann expression outside the dilute limit. Site exclusion and defect interactions then become important.
- Forgetting charge neutrality. The concentrations of charged defects and electronic carriers are coupled.
- Confusing total defects with mobile defects. Association and ordering can reduce the mobile fraction.
- Equating low formation energy with high conductivity. Migration barriers and microstructure also control transport.
- Ignoring phase stability constraints on chemical potentials. Arbitrary reservoir values may describe an unstable host phase.
15. Summary
Defect formation energy describes the thermodynamic cost of creating a point defect under specified chemical and electronic conditions. At finite temperature, equilibrium concentrations are governed by formation free energies and configurational statistics. In the dilute independent-defect limit,
[ [D]N (-) ]
For charged defects, formation energies depend on the Fermi level, while chemical potentials determine the energetic cost of exchanging atoms with the environment. The equilibrium Fermi level and defect populations must be found self-consistently using charge neutrality, site balance, and relevant mass-action relations.
In functional oxides, this framework connects atomistic defect calculations to oxygen nonstoichiometry, dopant compensation, proton uptake, and transport. The most reliable predictions combine defect energetics with finite-temperature thermodynamics and experimental validation.
Further reading
C. Freysoldt, B. Grabowski, T. Hickel, J. Neugebauer, G. K. H. Madsen, D. Neuhauser, and A. Janotti, “First-principles calculations for point defects in solids,” Reviews of Modern Physics 86, 253 (2014).
S. B. Zhang and J. E. Northrup, “Chemical potential dependence of defect formation energies in GaAs: Application to Ga self-diffusion,” Physical Review Letters 67, 2339 (1991).
A. R. West, Solid State Chemistry and its Applications, 2nd ed., Wiley.
D. R. Gaskell and D. E. Laughlin, Introduction to the Thermodynamics of Materials, 6th ed., CRC Press.