For the negative-sign equation, take a nonzero separated solution . Separation of variables gives and . The zero Dirichlet boundary conditions on force with a positive integer and , up to a constant. Zero boundary values on similarly require with a positive integer . A zero or negative separation eigenvalue would give linear or hyperbolic factors, whose two zero endpoint values force the trivial solution.
The only positive integer pairs with are and . By the complete Dirichlet eigensystem of a rectangle, these span the whole eigenspace, so
For the positive-sign equation there is only the zero solution. One can use the same strictly positive eigenvalues of , or directly apply the zero Dirichlet solution of the modified Helmholtz equation. Multiplying by and using Green's first identity gives
where . Both nonnegative integrals must vanish, proving .
Seek a separated nonzero solution . Division by on intervals where the factors are nonzero gives . Write and , with both factors zero at their endpoints .
For , a nonpositive gives only the zero solution under the two endpoint conditions: linear functions vanish identically when , and the hyperbolic solutions do likewise for . For , the endpoints force with positive integer and . Applying the same argument to gives and , with positive integer .
The only positive integer pairs satisfying are and . By superposition,
To establish that this is the most general sufficiently regular solution, expand in the complete product Fourier sine series on the square. Substitution, or integration by parts against each sine mode, gives for every coefficient. Hence only the two displayed modes remain; all boundary terms vanish because of the given zero values. This is the eigenvalue- case of the Dirichlet eigensystem of a rectangle. The zero solution is included when .
Figure 1.
The two independent zero-boundary modes sharing Laplacian eigenvalue 25 on the square
. Details.