Separation of Variables Method: Solving PDEs Step by Step

The separation of variables method solves partial differential equations by guessing that the solution is a product of single-variable functions, for example u(x, t) = X(x) times T(t). When this guess is substituted into the PDE, the equation splits into ordinary differential equations that are far easier to solve.

This matters because the heat, wave, and Laplace equations are almost always solved this way. The method turns one hard problem in two variables into two manageable problems in one variable each, and it reveals the natural vibration modes of the system.

Key Concepts

• Product ansatz: the trial solution u(x, t) = X(x) T(t), where X depends only on position and T only on time.
• Separation constant: after dividing by XT, each side depends on a different variable, so both must equal the same constant, usually written as minus lambda.
• Eigenvalue problem: the spatial equation X double-prime + lambda X = 0 together with boundary conditions allows solutions only for special values lambda(n), the eigenvalues.
• Eigenfunctions: the special spatial shapes, such as sin(n pi x / L), belonging to each eigenvalue.

Formulas

• Heat equation separation: X double-prime / X = T prime / (alpha T) = minus lambda, giving X double-prime + lambda X = 0 and T prime + alpha lambda T = 0.
• Wave equation separation: X double-prime / X = T double-prime / (c squared T) = minus lambda, giving X double-prime + lambda X = 0 and T double-prime + lambda c squared T = 0.
• For a rod of length L with zero temperature at both ends, the eigenvalues are lambda(n) = (n pi / L) squared for n = 1, 2, 3 and so on.
• Time decay of the n-th heat mode: T(n)(t) = exp(minus alpha (n pi / L) squared t), so higher modes die out faster.

Worked Example

A rod of length L = 1 meter has thermal diffusivity alpha = 0.1 square meters per second, ends held at zero temperature, and initial profile u(x, 0) = sin(pi x). Find how the temperature decays.

1. The initial profile already matches the n = 1 mode sin(pi x), so only B1 = 1 is nonzero.
2. The decay constant for n = 1 is alpha times (pi / L) squared = 0.1 times 9.8696 = 0.98696 per second.
3. The characteristic decay time is tau = 1 / 0.98696, about 1.013 seconds.
4. The solution is u(x, t) = sin(pi x) times e to the power (minus t / 1.013).
5. After 1 second, the amplitude factor is e to the power (minus 0.987), about 0.373, so the rod keeps only 37 percent of its initial temperature bump.

Common Misconceptions

• "Separation of variables works for any PDE." It works for linear PDEs in simple geometries with compatible boundary conditions. Nonlinear equations and irregular shapes usually defeat it.
• "One product solution is the answer." A single product X(x)T(t) almost never satisfies the initial condition. The physical solution is the infinite sum of modes.
• "The separation constant can be anything." Boundary conditions force its sign: for a rod with zero ends, only negative separation constants give fitting sine modes.

Real-World Applications

• Cooling of metal castings and food: each thermal mode decays with its own time constant, and engineers keep only the slowest few.
• Standing waves on strings and in organ pipes are the separated modes of the wave equation.

Separation of variables converts a PDE into ordinary differential equations plus an eigenvalue problem, and the sum of modes builds the full solution. Those textbook sine modes for strings and rods come from exactly this machinery.