Configuration State Functions (CSFs)¶

Once the active space has been selected, the next step in a CASSCF calculation is to generate all possible electronic configurations within that active space.

Rather than working directly with Slater determinants, GAMESS constructs Configuration State Functions (CSFs). These are spin-adapted combinations of Slater determinants that possess the correct spin symmetry.

The complete CASSCF wavefunction is ultimately expressed as a linear combination of these CSFs.

Why Not Use a Single Determinant?¶

Recall that RHF assumes a single electronic configuration.

Wavefunction

↓

One Slater Determinant

This works well for molecules whose electronic structure is dominated by one configuration.

However, many molecules require several important configurations.

Instead of

Ψ = Configuration 1

we require

Ψ = Configuration 1
   + Configuration 2
   + Configuration 3
   + ...

The question is:

How do we generate these configurations?

The answer is through Configuration State Functions.

From Active Space to Configurations¶

Suppose we have selected a

CAS(2,2)

active space.

This means

2 active electrons
2 active orbitals

The two electrons may occupy these orbitals in several different ways.

For example,

Configuration 1

Orbital A   ↑↓

Orbital B   __

or

Configuration 2

Orbital A   ↑

Orbital B   ↓

or

Configuration 3

Orbital A   __

Orbital B   ↑↓

Each represents a different electronic arrangement.

The Problem with Slater Determinants¶

At first glance, we might think each configuration can simply be represented by a Slater determinant.

Unfortunately, this introduces a problem.

Consider two singly occupied orbitals.

Orbital A   ↑

Orbital B   ↓

The determinant

|↑  ↓|

does not possess a well-defined total spin.

Similarly,

|↓  ↑|

also lacks pure spin symmetry.

Neither determinant is an eigenfunction of the total spin operator.

This is unacceptable because molecular electronic states must have a definite spin multiplicity.

Spin Adaptation¶

To obtain wavefunctions with the correct spin symmetry, determinants are combined together.

For example,

instead of using

|↑  ↓|

or

|↓  ↑|

individually,

we construct

|↑  ↓⟩ + |↓  ↑⟩

or

|↑  ↓⟩ − |↓  ↑⟩

These combinations possess well-defined spin quantum numbers.

Such spin-adapted combinations are called Configuration State Functions (CSFs).

What Is a CSF?¶

A Configuration State Function is therefore

a spin-adapted electronic configuration,
constructed from one or more Slater determinants,
possessing a definite total spin,
and suitable for use in multiconfigurational calculations.

Instead of writing

Wavefunction

↓

Determinants

CASSCF uses

Wavefunction

↓

Configuration State Functions

Why Does GAMESS Use CSFs?¶

The GUGA algorithm implemented in GAMESS works directly with CSFs.

Using CSFs provides several important advantages.

Correct spin symmetry is preserved automatically.
The CI matrix becomes smaller than a determinant-based expansion.
Degenerate spin states are treated consistently.
Computational efficiency is improved.

This is why GAMESS employs the Graphical Unitary Group Approach (GUGA) rather than a determinant-based CI algorithm.

Number of CSFs¶

One important feature of CASSCF is that the number of CSFs grows rapidly with the size of the active space.

For example,

Active Space	Approximate Number of CSFs
CAS(2,2)	Few
CAS(4,4)	Tens
CAS(6,6)	Hundreds
CAS(8,8)	Thousands
CAS(10,10)	Tens of thousands

The computational cost therefore increases rapidly as the active space becomes larger.

Relationship Between Determinants and CSFs¶

The hierarchy can be summarized as

Atomic Orbitals

↓

Molecular Orbitals

↓

Active Space

↓

Slater Determinants

↓

Configuration State Functions

↓

CASSCF Wavefunction

Notice that CSFs are built from Slater determinants, and the wavefunction is built from CSFs.

Example¶

Suppose our active space contains four orbitals and four electrons.

One possible occupation is

Another is

Each occupation pattern represents a different electronic configuration.

For every allowed occupation,

GAMESS constructs the corresponding spin-adapted Configuration State Function.

These CSFs later become the basis functions used to build the CASSCF wavefunction.

Why Are CSFs Important?¶

Without CSFs,

spin symmetry would not be preserved,
excited states could mix incorrectly,
CI calculations would become unnecessarily large.

CSFs provide a mathematically rigorous and computationally efficient basis for describing multiconfigurational electronic states.

They form the foundation of every CASSCF calculation performed by GAMESS.

Looking Ahead¶

Once all Configuration State Functions have been generated, the next step is to combine them into a single multiconfigurational wavefunction.

Each CSF contributes with a different weight, known as its Configuration Interaction (CI) coefficient.

Determining these coefficients is one of the primary goals of the CASSCF calculation.

The next chapter introduces the mathematical form of the CASSCF wavefunction.

Key Takeaways¶

A Configuration State Function (CSF) is a spin-adapted electronic configuration.
CSFs are constructed from one or more Slater determinants.
GAMESS generates all allowed CSFs within the chosen active space.
Using CSFs guarantees the correct spin symmetry of the electronic wavefunction.
The complete CASSCF wavefunction is expressed as a linear combination of CSFs.
The number of CSFs grows rapidly with the size of the active space.