The Eigenvalue Problem¶

After constructing the mass-weighted Hessian matrix, Gaussian must determine the natural vibrational motions of the molecule. This is achieved by solving an eigenvalue problem, one of the most important mathematical operations in molecular vibrational analysis.

The eigenvalue problem transforms the coupled motions of all atoms into a set of independent normal modes, each associated with a unique vibrational frequency.

Why is an Eigenvalue Problem Needed?¶

The Hessian matrix contains information about

the strength of every chemical bond,
the coupling between atomic motions,
the curvature of the Potential Energy Surface.

However, these quantities are mixed together.

To identify the independent molecular vibrations, the Hessian must be transformed into a coordinate system where each vibration is completely independent.

This transformation is accomplished through matrix diagonalization.

Matrix Diagonalization¶

The mass-weighted Hessian can be represented as

\[ \mathbf{F} \]

where F is the mass-weighted force constant matrix.

Gaussian diagonalizes this matrix by solving

\[ \boxed{ \mathbf{F}\mathbf{L} = \mathbf{L}\mathbf{\Lambda} } \]

where

F is the mass-weighted Hessian matrix,
L contains the eigenvectors,
Λ is a diagonal matrix containing the eigenvalues.

After diagonalization,

all coupling between vibrational motions disappears.

Each diagonal element now represents one independent vibration.

The Eigenvalue Equation¶

More commonly, the eigenvalue problem is written as

\[ \boxed{ \mathbf{F}\mathbf{x} = \lambda\mathbf{x} } \]

where

F is the mass-weighted Hessian,
x is an eigenvector,
λ is an eigenvalue.

Gaussian solves this equation for every vibrational mode of the molecule.

What is an Eigenvector?¶

An eigenvector describes how the atoms move during a particular vibration.

For one vibrational mode,

the eigenvector contains

the direction of motion,
the relative displacement,
the contribution of every atom.

For example,

O ←──── C ────→ O

and

O ↑
 \
  C
 /
O ↓

represent two different eigenvectors corresponding to different normal modes.

GaussView uses these eigenvectors to animate molecular vibrations.

What is an Eigenvalue?¶

The eigenvalue determines how fast the corresponding vibration occurs.

Large eigenvalues indicate

steep potential wells,
stiff bonds,
high vibrational frequencies.

Small eigenvalues correspond to

soft molecular motions,
weaker restoring forces,
lower frequencies.

Thus,

each eigenvalue is directly related to one vibrational frequency.

From Eigenvalues to Frequencies¶

Once the eigenvalues have been obtained,

Gaussian converts them into vibrational frequencies using

\[ \boxed{ \nu_i = \frac{1}{2\pi c} \sqrt{\lambda_i} } \]

where

\( \nu_i \) is the vibrational frequency,
\( c \) is the speed of light,
\( \lambda_i \) is the corresponding eigenvalue.

The frequencies reported in the Gaussian output are expressed in cm⁻¹.

Positive and Negative Eigenvalues¶

The sign of the eigenvalue provides important information about the nature of the stationary point.

Positive Eigenvalue¶

Potential Energy

      ●
    /   \

A positive eigenvalue corresponds to

upward curvature,
a stable vibration,
a real vibrational frequency.

Negative Eigenvalue¶

Potential Energy

\        /
 \______/

A negative eigenvalue indicates

downward curvature,
an unstable direction,
an imaginary vibrational frequency.

This is the signature of a transition state.

Relationship Between Eigenvalues and Frequencies¶

The correspondence is straightforward.

Eigenvalue	Frequency	Interpretation
Positive	Real	Stable vibration
Zero	Near 0 cm⁻¹	Translation or rotation
Negative	Imaginary	Transition state direction

This relationship allows Gaussian to determine whether the optimized geometry is a minimum or a saddle point on the Potential Energy Surface.

Overall Procedure¶

The sequence followed by Gaussian is

Mass-Weighted Hessian
          │
          ▼
Matrix Diagonalization
          │
          ▼
Eigenvalues
          │
          ├──────────────► Vibrational Frequencies
          │
          ▼
Eigenvectors
          │
          ▼
Normal Modes

The eigenvalues determine the frequencies,

while the eigenvectors determine the atomic motions associated with each vibration.

Computational Considerations¶

For a molecule containing N atoms, the mass-weighted Hessian has dimensions

\[ 3N \times 3N \]

Diagonalizing this matrix yields

3N eigenvalues, and
3N eigenvectors.

Among these,

three eigenvalues correspond to translations,
three (or two for linear molecules) correspond to rotations,
the remaining eigenvalues describe the molecular vibrational modes.

Summary¶

The eigenvalue problem is the mathematical step that transforms the mass-weighted Hessian into independent molecular vibrations. By diagonalizing the Hessian, Gaussian obtains a set of eigenvalues and eigenvectors. The eigenvalues determine the vibrational frequencies, while the eigenvectors define the corresponding normal modes of vibration. Positive eigenvalues produce real frequencies characteristic of stable structures, whereas negative eigenvalues give rise to imaginary frequencies that identify transition states. Solving the eigenvalue problem therefore connects the curvature of the Potential Energy Surface with the observable vibrational properties of a molecule.

Next Section¶

The next chapter explains Vibrational Frequencies, showing how the eigenvalues obtained from the Hessian are converted into the frequencies reported by Gaussian and how these frequencies are interpreted in molecular spectroscopy.