Definition and Mathematical Formulation
In linear algebra, an eigenvalue λ of a linear operator A acting on a vector space V is a scalar that satisfies
\[ \mathbf{A}\,\mathbf{v}= \lambda \mathbf{v}, \]
for some non‑zero vector v ∈ V. The vector v is called an eigenvector (or eigenmode) associated with λ. The equation is often written as
\[ (\mathbf{A} - \lambda \mathbf{I})\mathbf{v}= \mathbf{0}, \]
where I is the identity operator. Non‑trivial solutions exist only when the determinant of \(\mathbf{A} - \lambda \mathbf{I}\) vanishes, giving the characteristic equation
\[ \det(\mathbf{A} - \lambda \mathbf{I}) = 0. \]
The roots λ of this polynomial are the eigenvalues; each eigenvalue may correspond to one or more linearly independent eigenvectors, forming an eigenspace. In physics, the operator A is typically a representation of a physical observable (e.g., Hamiltonian, angular momentum) or a dynamical matrix describing coupled degrees of freedom. The eigenvalues then have direct physical meaning—energies, frequencies, growth rates, etc.—while eigenvectors describe the corresponding stationary states or mode shapes.
Role in Quantum Mechanics
Quantum theory is fundamentally built upon eigenvalue problems. Observables are represented by Hermitian (self‑adjoint) operators \(\hat{O}\) on a Hilbert space. The spectral theorem guarantees that the eigenvalues of \(\hat{O}\) are real and that eigenvectors form a complete orthonormal basis (possibly generalized in the case of continuous spectra). The canonical example is the time‑independent Schrödinger equation
\[ \hat{H}\,\psi = E\,\psi, \]
where \(\hat{H}\) is the Hamiltonian operator, \(E\) an energy eigenvalue, and \(\psi\) the corresponding stationary state (eigenfunction). Solving this eigenvalue problem yields the quantized energy levels of bound systems (atoms, molecules, quantum wells) and the associated wavefunctions that determine all measurable properties.
Similarly, angular momentum operators satisfy
\[ \hat{J}^2\,|j,m\rangle = \hbar^{2}j(j+1)\,|j,m\rangle, \qquad \hat{J}_z\,|j,m\rangle = \hbar m\,|j,m\rangle, \]
with eigenvalues \(\hbar^{2}j(j+1)\) and \(\hbar m\). The discrete spectra of these operators underpin selection rules, spin statistics, and the structure of multiplets. In many‑body physics, the diagonalization of the Hamiltonian matrix (or its second‑quantized form) provides quasiparticle energies, band structures, and correlation functions. The eigenvectors encode the composition of many‑body states in terms of single‑particle bases.
Classical Mechanics and Vibrational Analysis
In classical mechanics, eigenvalue problems arise when linearizing the equations of motion about an equilibrium configuration. Consider a system of N coupled oscillators with generalized coordinates \(\mathbf{q}\). Near equilibrium, the dynamics are governed by
\[ \mathbf{M}\,\ddot{\mathbf{q}} + \mathbf{K}\,\mathbf{q}= \mathbf{0}, \]
where M is the mass matrix (positive definite) and K the stiffness (force‑constant) matrix. Seeking harmonic solutions \(\mathbf{q}(t)=\mathbf{u}\,e^{i\omega t}\) leads to
\[ (\mathbf{K} - \omega^{2}\mathbf{M})\mathbf{u}= \mathbf{0}. \]
Thus \(\omega^{2}\) are eigenvalues of the generalized eigenproblem \(\mathbf{K}\mathbf{u}= \omega^{2}\mathbf{M}\mathbf{u}\), and the eigenvectors \(\mathbf{u}\) describe the normal mode shapes. The eigenfrequencies \(\omega\) are observable as resonant frequencies in spectroscopy, structural engineering, and molecular vibration spectra (infrared and Raman).
Analogous formulations appear in fluid dynamics (e.g., stability of shear flows) and plasma physics, where linear perturbations satisfy eigenvalue equations whose growth rates determine stability thresholds. In all such cases, eigenvectors represent the spatial structure of the perturbation, while eigenvalues dictate temporal evolution.
Applications in Continuum Mechanics and Electromagnetism
Eigenvalue methods extend to continuous media described by partial differential equations (PDEs). In elasticity, the Navier‑Cauchy equations can be cast as an eigenvalue problem for wave propagation:
\[ \mu \nabla^{2}\mathbf{u} + (\lambda + \mu)\nabla(\nabla\!\cdot\!\mathbf{u}) = \rho\,\partial_{t}^{2}\mathbf{u}, \]
leading to longitudinal and transverse wave speeds derived from eigenvalues of the elastic modulus tensor. In electromagnetic theory, waveguide and cavity modes are obtained by solving the Helmholtz eigenvalue problem
\[ \nabla \times (\nabla \times \mathbf{E}) = \left(\frac{\omega}{c}\right)^{2}\varepsilon(\mathbf{r})\,\mathbf{E}, \]
with eigenvalues \((\omega/c)^{2}\) and eigenfunctions \(\mathbf{E}\) representing the field distribution. The orthogonality of modes follows from the self‑adjoint nature of the curl‑curl operator under appropriate boundary conditions.
In photonic crystals and metamaterials, band structures are computed from eigenvalue problems defined on periodic unit cells (Bloch's theorem). The resulting dispersion relations determine allowed and forbidden frequency ranges (photonic band gaps), which are central to designing waveguides, filters, and light‑trapping devices.
Computational Methods and Numerical Implementation
Exact analytical solutions of eigenvalue problems are rare beyond low‑dimensional or highly symmetric systems. Numerical linear algebra provides the primary tools for physicists. For dense matrices, direct methods such as the QR algorithm (or its shifted variants) compute all eigenvalues and eigenvectors with \(O(N^{3})\) scaling. Sparse or structured matrices—common in discretized PDEs—are treated with iterative techniques: the Lanczos algorithm for Hermitian problems, Arnoldi iteration for non‑Hermitian cases, and the implicitly restarted variants (IRL, IRAM) that target a few eigenpairs of interest.
Finite element, finite difference, and spectral methods convert differential operators into large matrices, preserving symmetry properties that guarantee real eigenvalues for physical observables. Preconditioning, matrix-free formulations, and parallelization are essential for large‑scale simulations (e.g., electronic structure calculations with density‑functional theory, or seismic wave propagation).
Convergence criteria rely on residual norms \(\|\mathbf{A}\mathbf{v} - \lambda \mathbf{v}\|\) and orthogonality checks. In quantum many‑body problems, the exponential growth of Hilbert space dimension forces the use of specialized techniques such as the Density Matrix Renormalization Group (DMRG) and tensor‑network methods, which effectively capture dominant eigenvectors (ground states) without full diagonalization.
Eigenvalue and eigenvector concepts constitute a unifying mathematical framework across physics. Whether describing quantized energy levels, vibrational normal modes, stability growth rates, or electromagnetic resonances, the eigenvalue problem translates the abstract action of linear operators into physically measurable quantities and mode shapes. Their pervasive presence underpins both theoretical insight and practical computation in modern scientific research.