Eigenvalues and Eigenfunctions — Question 4

Section 27-2 · Calculus pages 4747–4748

Show solutionHide solution

Question 4

Let h∈ℝh\in\mathbb R control the endpoint condition −y″=λy,0≤x≤1,y(0)=0,y′(1)=hy(1).-y''=\lambda y,\qquad 0\leq x\leq 1,\qquad y(0)=0,\quad y'(1)=h\,y(1). Unlike fixed zero endpoint values, this boundary condition can permit a negative eigenvalue.

Tasks

  1. Determine precisely when zero is an eigenvalue and identify its eigenspace.

  2. For λ=−μ2<0\lambda=-\mu^2<0, derive the equation for μ\mu. Prove that there is exactly one negative eigenvalue when h>1h>1 and none when h≤1h\leq 1.

  3. For λ=k2>0\lambda=k^2>0, derive the frequency equation without losing or adding modes by division. Locate all its roots by intervals between multiples of π\pi, including the special first interval.

  4. Derive the energy identity and explain why it does not always imply nonnegative eigenvalues. For h=2h=2, bound the negative eigenvalue using 1<μ<21<\mu<2, and sketch the equation locating μ\mu.

Question 4 – Solution

Strategy. Separate all three signs and prove root counts by monotonicity on the correct domains.

Step 1: Check the zero parameter directly. At λ=0\lambda=0, y=Axy=Ax and the right condition is A=hAA=hA. Thus zero is an eigenvalue exactly when h=1h=1, with eigenspace span⁡{x}\operatorname{span}\{x\}.

Step 2: Count the negative modes. For λ=−μ2\lambda=-\mu^2, the nonzero solution is a multiple of sinh⁡(μx)\sinh(\mu x), and its endpoint condition is μcoth⁡μ=h\boxed{\mu\coth\mu=h}. The function g(μ)=μcoth⁡μg(\mu)=\mu\coth\mu tends to one as μ↓0\mu\downarrow 0 and to infinity as μ→∞\mu\to\infty. Its derivative has numerator sinh⁡μcosh⁡μ−μ>0\sinh\mu\cosh\mu-\mu>0, since this numerator starts at zero and has derivative 2sinh⁡2μ>02\sinh^2\mu>0. Thus precisely h>1h>1 gives one negative mode.

Step 3: Locate every positive mode. Here y=Asin⁡(kx)y=A\sin(kx) and kcos⁡k=hsin⁡kk\cos k=h\sin k. A positive multiple of π\pi cannot solve this equation, so division by sin⁡k\sin k is safe at a root: kcot⁡k=h.\boxed{k\cot k=h.} Its derivative is (sin⁡kcos⁡k−k)/sin⁡2k<0(\sin k\cos k-k)/\sin^2k<0 for k>0k>0 away from poles. On each (nπ,(n+1)π)(n\pi,(n+1)\pi), n≥1n\geq 1, it decreases from +∞+\infty to −∞-\infty, giving exactly one root for every hh. On (0,π)(0,\pi) it decreases from the unattained limit one to −∞-\infty: one root if h<1h<1, none if h≥1h\geq 1.

Step 4: Retain the endpoint energy term. Integration by parts gives λ∫01y2dx=∫01(y′)2dx−hy(1)2\lambda\int_0^1y^2dx=\int_0^1(y')^2dx-hy(1)^2, which can be negative. For h=2h=2, g(1)<2<g(2)g(1)<2<g(2): the first inequality follows from e2>3e^2>3 and g(2)>2g(2)>2 follows from coth⁡2>1\coth 2>1. Thus 1<μ<21<\mu<2 and −4<λ−<−1\boxed{-4<\lambda_-<-1}. The graph locates the unique crossing.

Diagram shown in the original SVG worksheet.