Exchange–Correlation Functionals¶

In the previous chapter, we introduced the Kohn–Sham equations,

\[ \left[ -\frac12\nabla^2 + V_{\mathrm{eff}}(\mathbf r) \right] \phi_i = \varepsilon_i \phi_i, \]

where the effective potential contains three contributions,

\[ V_{\mathrm{eff}} = V_{ne} + V_H + V_{xc}. \]

Among these terms,

the electron–nucleus attraction (\(V_{ne}\)) is known exactly,
the classical Coulomb (Hartree) potential (\(V_H\)) is known exactly,
only the exchange–correlation potential (\(V_{xc}\)) is unknown.

The exchange–correlation functional is therefore the central approximation in Density Functional Theory and is largely responsible for the success—or failure—of a DFT calculation.

What is Exchange–Correlation Energy?¶

The total electronic energy in Kohn–Sham DFT is written as

\[ \boxed{ E[\rho] = T_s[\rho] + V_{ne}[\rho] + J[\rho] + E_{xc}[\rho] } \]

where

Term	Meaning
\(T_s\)	Kinetic energy of non-interacting electrons
\(V_{ne}\)	Electron–nucleus attraction
\(J\)	Classical electron–electron repulsion
\(E_{xc}\)	Exchange–correlation energy

The exchange–correlation term contains all remaining quantum mechanical effects that are not included in the first three terms.

What Does \(E_{xc}\) Contain?¶

Although represented by a single mathematical expression,

the exchange–correlation functional accounts for several physical effects simultaneously.

These include

Exchange interactions arising from the Pauli Exclusion Principle,
Dynamic electron correlation due to the instantaneous motion of electrons,
Corrections to the kinetic energy,
Many-body quantum effects that cannot be described by classical electrostatics.

In practice,

\[ E_{xc} = E_x + E_c, \]

where

\(E_x\) is the exchange energy,
\(E_c\) is the correlation energy.

Why Can't We Calculate It Exactly?¶

Unfortunately,

the exact mathematical expression for

\[ E_{xc}[\rho] \]

is unknown.

If it were known,

Density Functional Theory would produce exact ground-state energies for all molecules.

Instead,

chemists develop approximate exchange–correlation functionals that reproduce experimental measurements and high-level quantum chemical calculations as accurately as possible.

The quality of a DFT calculation therefore depends strongly on the functional that is chosen.

Families of Exchange–Correlation Functionals¶

Over the years, many approximations have been developed.

These functionals are often grouped according to their complexity and accuracy.

Local Density Approximation (LDA)¶

The simplest approximation is the Local Density Approximation (LDA).

It assumes that the exchange–correlation energy at each point depends only on the electron density at that point,

\[ E_{xc} = E_{xc}[\rho(\mathbf r)]. \]

LDA works surprisingly well for

metals,
crystalline solids,
nearly uniform electron densities.

However,

it is generally less accurate for molecular systems.

Generalized Gradient Approximation (GGA)¶

The Generalized Gradient Approximation (GGA) improves upon LDA by including information about how rapidly the electron density changes in space.

The functional therefore depends on

\[ \rho(\mathbf r) \quad\text{and}\quad \nabla\rho(\mathbf r). \]

Because molecular electron densities vary significantly around atoms and chemical bonds,

GGA functionals usually provide better molecular geometries and energies than LDA.

Common GGA functionals include

PBE,
BP86,
BLYP.

Hybrid Functionals¶

Hybrid functionals combine

Density Functional Theory exchange,
Hartree–Fock exact exchange.

Their general form is

\[ E_x = aE_x^{HF} + (1-a)E_x^{DFT}, \]

where \(a\) is a mixing coefficient.

Including a fraction of exact Hartree–Fock exchange often improves

thermochemistry,
reaction energies,
molecular geometries,
activation barriers.

One of the most successful hybrid functionals is B3LYP, which will be discussed in the next chapter.

Jacob's Ladder of DFT¶

Exchange–correlation functionals are often organized using Jacob's Ladder, a conceptual framework introduced by John Perdew.

Each rung incorporates additional physical information and generally improves accuracy.

                 Chemical Accuracy
                       ▲

            Double Hybrids
                  ▲

           Hybrid Functionals
                  ▲

      Meta-Generalized Gradient
            Approximation
                  ▲

 Generalized Gradient Approximation
                  ▲

     Local Density Approximation

                  ▲
             Hartree Model

As one climbs the ladder,

the functionals become

more accurate,
more computationally demanding.

Examples of Common Functionals¶

Some widely used exchange–correlation functionals are listed below.

Functional	Type	Typical Applications
LDA	Local Density	Solids, metals
PBE	GGA	General chemistry, materials
BLYP	GGA	Organic molecules
B3LYP	Hybrid	General molecular chemistry
PBE0	Hybrid	Thermochemistry
M06-2X	Meta-hybrid	Main-group chemistry
ωB97X-D	Range-separated hybrid	Non-covalent interactions

Each functional has strengths and limitations, and no single functional is universally optimal.

Which Functional Does Gaussian Use?¶

Gaussian supports hundreds of exchange–correlation functionals.

The functional is specified directly in the route section of the input file.

For example,

# B3LYP/6-31G

tells Gaussian to use

the B3LYP hybrid functional,
together with the 6-31G basis set.

Changing only the functional,

# PBE0/6-31G

or

# M062X/6-31G

changes how the exchange–correlation energy is evaluated while leaving the remainder of the calculation unchanged.

Why is B3LYP So Popular?¶

Among the many available functionals,

B3LYP has become one of the most widely used for molecular quantum chemistry because it provides a good balance between

computational efficiency,
accuracy,
numerical stability.

For decades,

B3LYP has been one of the standard choices for

molecular geometries,
vibrational frequencies,
reaction energies,
organic and inorganic molecules.

Although newer functionals are available, B3LYP remains an excellent introductory functional and is therefore used throughout this documentation.

Summary¶

The exchange–correlation functional is the only approximate component of Kohn–Sham Density Functional Theory. It accounts for exchange interactions, electron correlation, and other many-body quantum effects that cannot be described by classical electrostatics alone. Because its exact mathematical form is unknown, practical DFT calculations rely on approximate functionals such as LDA, GGA, and hybrid functionals. The choice of functional has a major influence on the accuracy of a calculation, making it one of the most important decisions in computational chemistry.

Next Section¶

The next chapter introduces the B3LYP Functional, one of the most widely used hybrid density functionals. We will examine its mathematical form, explain the meaning of its individual components, and discuss why it has become the default choice for many Gaussian calculations.