Small Oscillations and Normal Modes

Understanding how complex systems vibrate is fundamental to fields ranging from molecular chemistry to mechanical engineering. By learning to analyze small oscillations—motion near stable points—and their normal modes—the independent, synchronous patterns of vibration—you unlock the ability to predict and control the behavior of intricate coupled systems, from the bonds in a molecule to the springs in a car's suspension.

The Foundation: Linearizing Near Stable Equilibrium

The study of small oscillations begins with a stable equilibrium point, a configuration where the net force on a system is zero, and any small displacement results in a restoring force pulling it back. Think of a marble resting at the very bottom of a bowl; a gentle nudge causes it to roll back and forth around the bottom.

For any conservative system with potential energy $U(\mathbf{q})$, where $\mathbf{q}$ represents the generalized coordinates (e.g., positions, angles), an equilibrium point ${\mathbf{q}}_0$ satisfies the condition that the gradient of the potential vanishes: $\partial U/\partial q_i=0$ for all coordinates. To analyze motion near this point, we perform a Taylor expansion of the potential energy around ${\mathbf{q}}_0$. For small displacements ${\eta }_i=q_i-q_{0i}$, we keep terms up to second order:

$$U(\mathbf{q})\approx U({\mathbf{q}}_0)+\sum _i\frac{\partial U}{\partial q_i}{|}_0{\eta }_i+\frac{1}{2}\sum _{i,j}\frac{{\partial }^{2}U}{\partial q_i\partial q_j}{|}_0{\eta }_i{\eta }_j.$$

The first term is a constant we can ignore, and the first derivatives are zero at equilibrium. Defining the constant matrix elements $K_{ij}={\partial }^{2}U/\partial q_i\partial q_j{|}_0$, the potential simplifies to a quadratic form: $U\approx \frac{1}{2}{\sum }_{i,j}{K}_{ij}{\eta }_i{\eta }_j$. The matrix $\mathbf{K}$ is the force constant matrix (or Hessian matrix). Crucially, if the equilibrium is stable, $\mathbf{K}$ is a symmetric, positive-definite matrix.

Simultaneously, for small velocities, the kinetic energy $T$ is also a quadratic function of the generalized velocities ${\dot{\eta }}_i$. It can be written as $T\approx \frac{1}{2}{\sum }_{i,j}{M}_{ij}{\dot{\eta }}_i{\dot{\eta }}_j$, where $\mathbf{M}$ is the mass matrix (or inertia matrix), which is also symmetric and positive-definite. With both energy expressions quadratic, the resulting equations of motion derived from the Lagrangian ($L=T-U$) become a set of coupled linear differential equations:

$$\sum _j{M}_{ij}{\ddot{\eta }}_j=-\sum _j{K}_{ij}{\eta }_j,\text{or in matrix form:}\mathbf{M}\ddot{\mathbf{\eta }}=-\mathbf{K}\mathbf{\eta }.$$

This linearization process is the gateway to solving the dynamics of complex systems.

The Eigenvalue Problem and Normal Mode Frequencies

The coupled linear equations $\mathbf{M}\ddot{\mathbf{\eta }}=-\mathbf{K}\mathbf{\eta }$ admit simple harmonic solutions. We assume all coordinates oscillate with the same frequency $\omega$ and phase, expressed as $\mathbf{\eta }(t)=\mathbf{a}{e}^{i\omega t}$, where $\mathbf{a}$ is a constant amplitude vector. Substituting this trial solution yields:

$$-{\omega }^2\mathbf{Ma}=-\mathbf{Ka}\Rightarrow (\mathbf{K}-{\omega }^2\mathbf{M})\mathbf{a}=0.$$

This is a generalized eigenvalue problem. For non-trivial solutions (where $\mathbf{a}\ne 0$), the determinant must vanish: $\det (\mathbf{K}-{\omega }^2\mathbf{M})=0$. Solving this characteristic equation gives the normal mode frequencies ${\omega }_{\alpha }$ (where $\alpha =1,2,...,N$ for a system with $N$ degrees of freedom).

For each eigenvalue ${\omega }_{\alpha }^2$, we solve for the corresponding eigenvector ${\mathbf{a}}^{(\alpha )}$. This vector defines the normal mode pattern—the specific ratio of displacements for each coordinate during that pure mode of oscillation. In a given normal mode, all parts of the system move sinusoidally with the same frequency ${\omega }_{\alpha }$ and fixed phase relationship (either in-phase or exactly out-of-phase). The general motion of the system is then a superposition of all normal modes: $\mathbf{\eta }(t)={\sum }_{\alpha }{C}_{\alpha }{\mathbf{a}}^{(\alpha )}\cos ({\omega }_{\alpha }t+{\varphi }_{\alpha })$, where constants $C_{\alpha }$ and ${\varphi }_{\alpha }$ are determined by initial conditions.

Solving a Prototype System: Two Coupled Pendulums

Consider two identical simple pendulums of length $l$ and mass $m$, connected by a weak spring with constant $k$ attached halfway down their rods. Let the small angular displacements be ${\theta }_1$ and ${\theta }_2$. The potential energy has two contributions: gravitational and spring. Expanding for small angles ($\cos \theta \approx 1-{\theta }^2/2$), the gravitational potential gives terms like $mgl{\theta }^2/2$. The spring potential is $\frac{1}{2}k(x_1-x_2{)}^2\approx \frac{1}{2}k(l/2{)}^2({\theta }_1-{\theta }_2{)}^2$.

After constructing the Lagrangian, the mass matrix $\mathbf{M}$ and force constant matrix $\mathbf{K}$ for the coordinate vector $\mathbf{\eta }=({\theta }_1,{\theta }_2{)}^T$ are:

$$\mathbf{M}=m{l}^2\left(\begin{array}{cc}1& 0\\ 0& 1\end{array}\right),\mathbf{K}=\left(\begin{array}{cc}mgl+\frac{k{l}^2}{4}& -\frac{k{l}^2}{4}\\ -\frac{k{l}^2}{4}& mgl+\frac{k{l}^2}{4}\end{array}\right).$$

Solving the eigenvalue problem $(\mathbf{K}-{\omega }^2\mathbf{M})\mathbf{a}=0$ yields two normal modes:

1. In-phase mode (${\omega }_1$): ${\mathbf{a}}^{(1)}=(1,1)$. Both pendulums swing together. The spring is neither stretched nor compressed, so the frequency is simply that of a simple pendulum: ${\omega }_1^2=g/l$.
2. Out-of-phase mode (${\omega }_2$): ${\mathbf{a}}^{(2)}=(1,-1)$. The pendulums swing opposite each other. The spring is stretched and compressed, adding an extra restoring force: ${\omega }_2^2=g/l+k/(2m)$.

Any arbitrary initial motion of the two pendulums can be expressed as a combination of these two simple, independent patterns.

Applications in Physics and Chemistry

The formalism of small oscillations and normal modes is extraordinarily powerful and finds direct application across the sciences.

In molecular vibrations, the atoms in a molecule are the masses and the chemical bonds act as springs. A molecule with $N$ atoms has $3N$ degrees of freedom. After subtracting translations and rotations, the remaining $(3N-6)$ (or $(3N-5)$ for linear molecules) are vibrational normal modes. Solving the eigenvalue problem for the molecular force field ($\mathbf{K}$ matrix) predicts the infrared and Raman spectrum of the molecule, as each normal mode frequency corresponds to a possible absorption line.

In crystal lattice dynamics, atoms are arranged in a periodic structure. The coupled oscillator analysis is performed on a massive scale, considering interactions between many atoms in a unit cell and their neighbors. The resulting eigenvalue problem yields dispersion relations—plots of frequency versus wavevector—that describe how vibrations (phonons) propagate through the solid. These phonons determine crucial material properties like thermal conductivity, specific heat, and sound speed.

The principles also apply directly to coupled pendulum systems, electrical LC circuits, acoustics in rooms, and the stability analysis of mechanical structures. The core idea remains: complex, seemingly chaotic motion can be decomposed into simpler, fundamental oscillations that are intrinsic to the system's geometry and physical parameters.

Common Pitfalls

1. Linearizing Unstable Equilibria: The entire framework assumes a stable equilibrium, where the force constant matrix $\mathbf{K}$ is positive-definite. Attempting to apply it near an unstable point (like a marble balanced on top of an inverted bowl) will yield imaginary frequencies (${\omega }^2<0$), corresponding to exponential runaway solutions, not oscillation. Always verify stability first.
2. Ignoring the Mass Matrix: In problems with non-Cartesian coordinates or coupled inertias, the mass matrix $\mathbf{M}$ is not simply diagonal with masses. Using an incorrect $\mathbf{M}$ (e.g., forgetting moment of inertia in rotational problems) will lead to wrong eigenvalues. Always derive the kinetic energy correctly from first principles.
3. Confusing Normal Mode Amplitudes: The relative amplitudes in an eigenvector ${\mathbf{a}}^{(\alpha )}$ are fixed, but their absolute scale is not. The ratio $a_1^{(\alpha )}:a_2^{(\alpha )}:...$ is what defines the mode pattern. The overall amplitude of a mode in a given motion is set by the initial-condition constants $C_{\alpha }$.
4. Applying Beyond Small Amplitudes: The solutions are only valid for "small" displacements where the quadratic approximations for $T$ and $U$ hold. For large amplitudes, nonlinear terms become significant, leading to phenomena like frequency dependence on amplitude and energy transfer between modes, which this linear analysis cannot describe.

Summary

• The analysis of small oscillations involves linearizing the equations of motion around a stable equilibrium point, resulting in a set of coupled linear differential equations expressible in matrix form as $\mathbf{M}\ddot{\mathbf{\eta }}=-\mathbf{K}\mathbf{\eta }$.
• Solving the associated generalized eigenvalue problem $(\mathbf{K}-{\omega }^2\mathbf{M})\mathbf{a}=0$ yields the system's normal mode frequencies ${\omega }_{\alpha }$ and their corresponding displacement patterns, the eigenvectors ${\mathbf{a}}^{(\alpha )}$.
• In a given normal mode, all parts of the system oscillate sinusoidally with the same frequency and a fixed phase relationship; any general motion is a superposition of these independent modes.
• This formalism is directly applicable to predicting the vibrational spectra of molecules and the phonon dispersion relations in crystal lattices, as well as to understanding a wide variety of mechanical and electrical oscillatory systems.
• Critical steps include verifying the stability of the equilibrium, correctly deriving both the mass ($\mathbf{M}$) and force constant ($\mathbf{K}$) matrices, and remembering that the linear solutions are only valid for small-amplitude displacements.