Chapter 8 · Functional Analysis

8.4 Compact Self-Adjoint Operators and the Spectral Theorem

Compact operators behave like matrices: they are limits of finite-rank operators, their nonzero spectrum consists of eigenvalues of finite multiplicity accumulating only at $0$, and — when they are self-adjoint — they are diagonal in an orthonormal basis of eigenvectors. Applications: integral equations and Sturm–Liouville problems.

Every real symmetric matrix can be diagonalised by an orthogonal matrix: there is an orthonormal basis of eigenvectors, the eigenvalues are real, and $\|A\|$ is the largest $|\lambda|$ (§3.3). Section 8.3 gave us orthonormal bases in infinite dimensions; the natural question is which operators on a Hilbert space can be diagonalised in the same way. Not all bounded self-adjoint operators can: multiplication by $t$ on $L^2[0,1]$ is self-adjoint but has no eigenvectors at all. What rescues the finite-dimensional picture is compactness — an operator that maps bounded sets to sets with compact closure. Integral operators with square-integrable kernels are compact, and for them Hilbert and Schmidt proved, in the first decade of the twentieth century, that a symmetric kernel is diagonalised by its eigenfunctions: $$Tx=\sum_n\lambda_n\langle x,e_n\rangle e_n,\qquad\lambda_n\to0.$$

This spectral theorem for compact self-adjoint operators is the engine behind the classical theory of integral equations (the Fredholm alternative), the completeness of the eigenfunctions of Sturm–Liouville problems (and therefore of Fourier sine series, Legendre and Bessel expansions), principal component analysis of random curves, and the normal modes of vibrating systems. GATE tests it through computations — the eigenvalues and norm of an integral operator with a "degenerate" kernel such as $x+t$, the spectrum of a diagonal or multiplication operator, the solvability of $u-\lambda Tu=f$ — and through MSQs on which operators are compact and what the spectrum of a compact operator can look like.

Compact Operators

📘 Definition 8.14 — Compact and finite-rank operators
Let $X,Y$ be normed spaces. A linear map $T:X\to Y$ is compact if the image $T(B)$ of the closed unit ball $B$ of $X$ has compact closure in $Y$; equivalently (in metric spaces), for every bounded sequence $(x_n)$ in $X$, the sequence $(Tx_n)$ has a convergent subsequence. $T$ has finite rank if $\dim T(X)\lt\infty$. We write $K(X,Y)$ for the compact operators and $K(X)=K(X,X)$. On a Hilbert space a finite-rank operator has the form $Tx=\sum_{i=1}^r\langle x,a_i\rangle b_i$.
📐 Theorem 8.29 — Basic properties of compact operators
Let $X,Y,Z$ be normed spaces.
  1. Every compact operator is bounded. Every bounded finite-rank operator is compact.
  2. $K(X,Y)$ is a subspace of $B(X,Y)$, and if $Y$ is a Banach space it is closed: a norm limit of compact operators is compact.
  3. Ideal property: if $T$ is compact and $S$ bounded, then $ST$ and $TS$ (when defined) are compact.
  4. Riesz's lemma: if $M$ is a closed proper subspace of $X$ and $0\lt\theta\lt1$, there is a unit vector $u$ with $\operatorname{dist}(u,M)\ge\theta$.
  5. The identity of $X$ is compact if and only if $\dim X\lt\infty$. Consequently a compact operator on an infinite-dimensional space has no bounded inverse.

Proof. (1) If $T$ were unbounded we could pick $\|x_n\|\le1$ with $\|Tx_n\|\ge n$; then $(Tx_n)$ has no convergent (indeed no bounded) subsequence. If $T$ is bounded with finite-dimensional range, $T(B)$ is a bounded subset of a finite-dimensional normed space, whose closure is compact (Heine–Borel, all norms being equivalent there — §8.1).

(2) Sums and scalar multiples: extract subsequences one after the other. For closedness let $\|T_n-T\|\to0$ with $T_n$ compact and let $(x_k)$ satisfy $\|x_k\|\le1$. Choose a subsequence on which $T_1x_k$ converges, a further subsequence on which $T_2x_k$ converges, and so on; the diagonal sequence $(z_k)$ makes every $(T_nz_k)_k$ convergent. Then $$\begin{aligned}\|Tz_k-Tz_j\|&\le\|(T-T_n)z_k\|+\|T_nz_k-T_nz_j\|\\&\quad+\|(T_n-T)z_j\|\\&\le2\|T-T_n\|+\|T_nz_k-T_nz_j\|.\end{aligned}$$ Given $\varepsilon$, fix $n$ with $2\|T-T_n\|\lt\varepsilon/2$, then $k,j$ large: $(Tz_k)$ is Cauchy and converges because $Y$ is complete.

(3) If $(x_n)$ is bounded, $(Tx_n)$ has a convergent subsequence and $S$ preserves convergence; and $(Sx_n)$ is bounded, so $(TSx_n)$ has a convergent subsequence.

(4) Pick $x\notin M$; $d=\operatorname{dist}(x,M)\gt0$ since $M$ is closed. Choose $m_0\in M$ with $\|x-m_0\|\le d/\theta$ and put $u=(x-m_0)/\|x-m_0\|$. For $m\in M$, $m_0+\|x-m_0\|m\in M$, so $$\|u-m\|=\frac{\big\|x-(m_0+\|x-m_0\|m)\big\|}{\|x-m_0\|}\ge\frac{d}{d/\theta}=\theta.$$

(5) If $\dim X\lt\infty$ the identity has finite rank. If $\dim X=\infty$, choose a unit vector $u_1$ and inductively, by (4) applied to the finite-dimensional (hence closed, §8.1) proper subspace $\operatorname{span}\{u_1,\dots,u_{n-1}\}$, a unit vector $u_n$ at distance $\ge\frac12$ from it. Then $\|u_n-u_m\|\ge\frac12$ for $m\ne n$: the unit ball is not compact. If $T$ were compact with $T^{-1}$ bounded, $I=T^{-1}T$ would be compact by (3). ∎

📘 Definition 8.15 — Hilbert–Schmidt operator
An operator $T\in B(H,K)$ between Hilbert spaces, $H$ separable, is a Hilbert–Schmidt operator if $\sum_n\|Te_n\|^2\lt\infty$ for some orthonormal basis $(e_n)$ of $H$; the Hilbert–Schmidt norm is $\|T\|_{HS}=\big(\sum_n\|Te_n\|^2\big)^{1/2}$. (By Parseval, for any orthonormal basis $(f_m)$ of $K$, $\sum_n\|Te_n\|^2=\sum_{n,m}|\langle Te_n,f_m\rangle|^2$ $=\sum_m\|T^*f_m\|^2$, so the value does not depend on the basis.)
📐 Theorem 8.30 — Compact operators on Hilbert spaces
Let $H,K$ be Hilbert spaces and $T\in B(H,K)$.
  1. If $T$ is compact, then $\|Te_n\|\to0$ for every orthonormal sequence $(e_n)$ in $H$.
  2. $T$ is compact if and only if it is a norm limit of finite-rank operators.
  3. $T$ is compact if and only if $T^*$ is compact.
  4. Every Hilbert–Schmidt operator is compact, and $\|T\|\le\|T\|_{HS}$. In particular an integral operator $(T_Ku)(x)=\int_a^bK(x,t)u(t)\,dt$ with $K\in L^2([a,b]^2)$ is compact on $L^2[a,b]$, with $$\begin{aligned}\|T_K\|&\le\|T_K\|_{HS}\\&=\Big(\int_a^b\!\!\int_a^b|K(x,t)|^2\,dt\,dx\Big)^{1/2}.\end{aligned}$$
  5. A diagonal operator $D_\lambda x=(\lambda_nx_n)$ on $\ell^2$ ($(\lambda_n)$ bounded) is compact if and only if $\lambda_n\to0$.

Proof. (1) If not, there are $\varepsilon\gt0$ and a subsequence with $\|Te_{n_k}\|\ge\varepsilon$; by compactness a further subsequence has $Te_{n_k}\to y$, and $\|y\|\ge\varepsilon$. But $\langle Te_{n_k},y\rangle=\langle e_{n_k},T^*y\rangle\to0$ by Bessel's inequality (§8.3), while it also tends to $\langle y,y\rangle\ne0$ — a contradiction.

(2) "If" is Theorem 8.29(1), (2). Conversely let $T$ be compact. The set $\overline{T(B)}$ is compact, hence separable, so $\overline{\operatorname{ran}T}$ is a separable Hilbert space with a countable orthonormal basis $(f_k)$ (§8.3). Let $P_n$ be the orthogonal projection onto $\operatorname{span}\{f_1,\dots,f_n\}$; $P_nT$ has finite rank. If $\|T-P_nT\|\not\to0$, there are $\varepsilon\gt0$ and unit vectors $x_n$ (along a subsequence) with $\|(I-P_n)Tx_n\|\ge\varepsilon$; passing to a further subsequence, $Tx_n\to y\in\overline{\operatorname{ran}T}$. Then $$\begin{aligned}\|(I-P_n)Tx_n\|&\le\|(I-P_n)(Tx_n-y)\|\\&\quad+\|(I-P_n)y\|\\&\le\|Tx_n-y\|+\|(I-P_n)y\|\to0,\end{aligned}$$ the last term by Parseval in $\overline{\operatorname{ran}T}$ — a contradiction.

(3) If $T_n=\sum_i\langle\cdot,a_i\rangle b_i$ has finite rank, so does $T_n^*=\sum_i\langle\cdot,b_i\rangle a_i$, and $\|T_n^*-T^*\|=\|T_n-T\|$ (§8.3). Apply (2) to $T$ and to $T^{**}=T$.

(4) Let $P_N$ project onto $\operatorname{span}\{e_1,\dots,e_N\}$. For $x=\sum c_ne_n$, by Cauchy–Schwarz in $\ell^2$, $$\begin{aligned}\|(T-TP_N)x\|&=\Big\|\sum_{n\gt N}c_nTe_n\Big\|\le\sum_{n\gt N}|c_n|\,\|Te_n\|\\&\le\|x\|\Big(\sum_{n\gt N}\|Te_n\|^2\Big)^{1/2}.\end{aligned}$$ So the finite-rank operators $TP_N$ converge to $T$ in norm, and $T$ is compact by (2); with $N=0$ the same estimate gives $\|T\|\le\|T\|_{HS}$. For the integral operator: for almost every $x$ the function $K_x=K(x,\cdot)$ is in $L^2$ (Fubini), and $(T_Ke_n)(x)=\langle K_x,\bar e_n\rangle$. Since $(\bar e_n)$ is again an orthonormal basis, Parseval gives $\sum_n|(T_Ke_n)(x)|^2=\|K_x\|^2$, and integrating in $x$ (monotone convergence) $\sum_n\|T_Ke_n\|^2=\int\|K_x\|^2dx=\|K\|_{L^2}^2$.

(5) If $\lambda_n\to0$, the truncations $D^{(N)}=D_{(\lambda_1,\dots,\lambda_N,0,\dots)}$ have finite rank and $\|D_\lambda-D^{(N)}\|=\sup_{n\gt N}|\lambda_n|\to0$. If $\lambda_n\not\to0$, then $\|D_\lambda e_n\|=|\lambda_n|\not\to0$, contradicting (1). ∎

✏️ Example 1 — Compact or not? Prove or disprove
Decide which of the following operators are compact: (a) $D x=(x_n/n)$ on $\ell^2$; (b) the right shift $S$ on $\ell^2$; (c) the Volterra operator $(Vu)(x)=\int_0^xu(t)\,dt$ on $L^2[0,1]$; (d) multiplication $(M_tu)(t)=t\,u(t)$ on $L^2[0,1]$; (e) the orthogonal projection onto an infinite-dimensional closed subspace; (f) $Tu(x)=\int_0^1(x+t)u(t)\,dt$ on $L^2[0,1]$.
(a) Compact Diagonal with $\lambda_n=\frac1n\to0$ (Theorem 8.30(5)); also Hilbert–Schmidt, $\|D\|_{HS}^2=\sum\frac1{n^2}=\frac{\pi^2}6$.
(b) Not compact $\|Se_n\|=\|e_{n+1}\|=1\not\to0$ along the orthonormal sequence $(e_n)$ (Theorem 8.30(1)). Any isometry on an infinite-dimensional space fails in the same way.
(c) Compact Kernel $K(x,t)=1$ for $t\le x$, $0$ otherwise: $\|K\|_{L^2}^2=\int_0^1x\,dx=\frac12$, so $V$ is Hilbert–Schmidt with $\|V\|\le\frac1{\sqrt2}$. (In fact $\|V\|=\frac2\pi\approx0.6366$ — GATE corner E4.)
(d) Not compact Let $u_n(t)=2\sin\big(2n\pi(t-\tfrac12)\big)$ on $[\tfrac12,1]$ and $u_n=0$ on $[0,\tfrac12)$: an orthonormal sequence supported where $t\ge\frac12$. Then $\|M_tu_n\|\ge\frac12\|u_n\|=\frac12\not\to0$.
(e) Not compact $P$ is the identity on its range, which contains an orthonormal sequence $(e_n)$ with $\|Pe_n\|=1$.
(f) Compact $Tu=x\langle u,1\rangle+\langle u,t\rangle$ has rank $2$.

Spectrum, Eigenvalues and Self-Adjoint Operators

In finite dimensions $\lambda$ is in the spectrum exactly when $A-\lambda I$ is not injective. In infinite dimensions an injective operator need not be onto, and the spectrum is defined through invertibility. Spectra are taken in complex Banach spaces; for self-adjoint operators everything below also works word for word over $\mathbb{R}$.

📘 Definition 8.16 — Resolvent set, spectrum, eigenvalues, spectral radius
Let $X$ be a complex Banach space and $T\in B(X)$. The resolvent set $\rho(T)$ is the set of $\lambda\in\mathbb{C}$ for which $T-\lambda I$ is a bijection of $X$ (its inverse is then bounded by the bounded inverse theorem, §8.2). The spectrum is $\sigma(T)=\mathbb{C}\setminus\rho(T)$. It contains the point spectrum $\sigma_p(T)$ of eigenvalues ($\ker(T-\lambda I)\ne\{0\}$) and the approximate eigenvalues: $\lambda$ with unit vectors $x_n$ such that $\|(T-\lambda)x_n\|\to0$. The spectral radius is $r(T)=\sup\{|\lambda|:\lambda\in\sigma(T)\}$.
📐 Theorem 8.31 — The spectrum of a bounded operator
Let $X$ be a complex Banach space and $T,A\in B(X)$.
  1. Neumann series. If $\|A\|\lt1$, then $I-A$ is invertible, $(I-A)^{-1}=\sum_{n\ge0}A^n$ and $\|(I-A)^{-1}\|\le\frac1{1-\|A\|}$.
  2. $\sigma(T)$ is a closed subset of the disc $\{|\lambda|\le\|T\|\}$, hence compact.
  3. $\sigma(T)\ne\varnothing$, and $r(T)=\lim_n\|T^n\|^{1/n}\le\|T\|$ (Gelfand's spectral radius formula).
  4. If $X$ is a Hilbert space, $\sigma(T^*)=\{\bar\lambda:\lambda\in\sigma(T)\}$.

Proof. (1) $\sum\|A^n\|\le\sum\|A\|^n\lt\infty$ and $B(X)$ is complete (§8.1), so $S=\sum A^n$ converges, with $\|S\|\le\frac1{1-\|A\|}$. The telescoping identities $(I-A)\sum_{n\le N}A^n=I-A^{N+1}=\sum_{n\le N}A^n(I-A)$ and $A^{N+1}\to0$ give $(I-A)S=S(I-A)=I$.

(2) If $|\lambda|\gt\|T\|$, then $T-\lambda I=-\lambda(I-T/\lambda)$ is invertible by (1). If $\lambda_0\in\rho(T)$ and $R=(T-\lambda_0)^{-1}$, then $$T-\lambda=(T-\lambda_0)\big(I-(\lambda-\lambda_0)R\big)$$ is invertible for $|\lambda-\lambda_0|\lt1/\|R\|$; so $\rho(T)$ is open and $\sigma(T)$ closed.

(3) (Sketch.) For $f\in B(X)^*$ the function $\lambda\mapsto f\big((T-\lambda)^{-1}\big)$ is holomorphic on $\rho(T)$ (expand $(T-\lambda)^{-1}$ in a Neumann series around each $\lambda_0$, as in (2)) and tends to $0$ as $|\lambda|\to\infty$, because $\|(T-\lambda)^{-1}\|\le\frac1{|\lambda|-\|T\|}$. If $\sigma(T)$ were empty it would be a bounded entire function, hence $0$ by Liouville's theorem (§5.1), for every $f$; by Hahn–Banach (§8.1) $(T-\lambda)^{-1}=0$, which is absurd. The formula for $r(T)$ comes from the Laurent expansion $(T-\lambda)^{-1}=-\sum_n T^n/\lambda^{n+1}$, whose radius of convergence is governed by $\limsup\|T^n\|^{1/n}$; we use only the statement. (4) $(T-\lambda)^*=T^*-\bar\lambda$, and an operator is invertible iff its adjoint is ($(S^{-1})^*=(S^*)^{-1}$). ∎

✏️ Example 2 — Four model spectra
Find the spectrum and the eigenvalues of (a) a diagonal operator $D_\lambda$ on $\ell^2$, in particular $\lambda_n=\frac1n$; (b) $M_t$ on $L^2[0,1]$; (c) the left shift $L$ and the right shift $S$ on $\ell^2$; (d) the Volterra operator $V$ on $L^2[0,1]$.
(a) Diagonal $D_\lambda e_n=\lambda_ne_n$, so every $\lambda_n$ is an eigenvalue; if $\mu\ne\lambda_n$ for all $n$, then $(D_\lambda-\mu)x=0$ forces $x=0$. If $\delta=\inf_n|\lambda_n-\mu|\gt0$, the inverse is the diagonal operator $(1/(\lambda_n-\mu))$, of norm $1/\delta$. If $\mu=\lim\lambda_{n_k}$, then $\|(D_\lambda-\mu)e_{n_k}\|\to0$: an approximate eigenvalue, not in $\rho$. So $\sigma(D_\lambda)=\overline{\{\lambda_n\}}$ and $\sigma_p=\{\lambda_n\}$. For $\lambda_n=\frac1n$: $\sigma=\{0\}\cup\{\frac1n\}$, and $0$ is in the spectrum but is not an eigenvalue.
(b) Multiplication For $\mu\notin[0,1]$, $1/(t-\mu)$ is bounded on $[0,1]$ and $M_{1/(t-\mu)}$ inverts $M_t-\mu$. For $\mu\in[0,1]$, $u_n=\sqrt n\,\mathbf 1_{I_n}$ with $I_n\subseteq[0,1]$ an interval of length $\frac1n$ containing $\mu$ has $\|u_n\|=1$ and $\|(M_t-\mu)u_n\|\le\frac1n$. So $\sigma(M_t)=[0,1]$. No eigenvalues: $(t-\mu)u(t)=0$ a.e. forces $u=0$ a.e.
(c) Shifts $Lx=\mu x$ means $x_{n+1}=\mu x_n$, i.e. $x=x_1(1,\mu,\mu^2,\dots)$, which is in $\ell^2$ iff $|\mu|\lt1$: $\sigma_p(L)$ is the open unit disc. As $\sigma(L)$ is closed and lies in $\{|\mu|\le\|L\|=1\}$, $\sigma(L)$ is the closed unit disc. Since $S=L^*$, Theorem 8.31(4) gives $\sigma(S)=$ the closed unit disc too, but $S$ has no eigenvalues: $Sx=\mu x$ reads $0=\mu x_1$, $x_1=\mu x_2$, $x_2=\mu x_3,\dots$, and both $\mu=0$ and $\mu\ne0$ force $x=0$.
(d) Volterra $V$ is compact (Example 1(c)), so $0\in\sigma(V)$ (Theorem 8.29(5)). $Vu=0$ gives $\int_0^xu=0$ for all $x$, so $u=0$; and $Vu=\mu u$ with $\mu\ne0$ makes $u=\mu^{-1}Vu$ continuous, then $C^1$, with $\mu u'=u$ and $u(0)=0$, so $u=0$. Moreover $\|V^n\|\le\frac1{n!}$ (Exercise II.5), hence $r(V)=\lim\|V^n\|^{1/n}=0$ and $\sigma(V)=\{0\}$ — although $\|V\|=\frac2\pi\gt0$. A compact operator can have no eigenvalues whatsoever.
📘 Definition 8.17 — Rayleigh quotient, numerical bounds, positive operators
Let $T\in B(H)$ be self-adjoint. Its Rayleigh quotient is $R(x)=\langle Tx,x\rangle/\|x\|^2$ ($x\ne0$), a real number, and its numerical bounds are $$m=\inf_{\|x\|=1}\langle Tx,x\rangle,\qquad M=\sup_{\|x\|=1}\langle Tx,x\rangle.$$ $T$ is positive, $T\ge0$, if $\langle Tx,x\rangle\ge0$ for all $x$. (On a complex Hilbert space, $\langle Tx,x\rangle\in\mathbb{R}$ for all $x$ already forces $T=T^*$, by polarization.)
📐 Theorem 8.32 — Self-adjoint operators
Let $T\in B(H)$ be self-adjoint with numerical bounds $m\le M$.
  1. Every eigenvalue of $T$ is real, and eigenvectors for different eigenvalues are orthogonal.
  2. $\|T\|=\sup_{\|x\|=1}|\langle Tx,x\rangle|=\max(|m|,|M|)$.
  3. $\sigma(T)\subseteq[m,M]$, and $m,M\in\sigma(T)$ (as approximate eigenvalues). Hence $\sigma(T)$ is real and $r(T)=\|T\|$.
  4. If a subspace $N$ is invariant under $T$ ($TN\subseteq N$), so is $N^\perp$.

Proof. (1) If $Tx=\lambda x$ with $\|x\|=1$, then $\lambda=\langle Tx,x\rangle=\langle x,Tx\rangle=\bar\lambda$. If also $Ty=\mu y$, $\mu\ne\lambda$, then $\lambda\langle x,y\rangle=\langle Tx,y\rangle=\langle x,Ty\rangle=\mu\langle x,y\rangle$, so $\langle x,y\rangle=0$.

(2) Let $\alpha=\sup_{\|x\|=1}|\langle Tx,x\rangle|$; clearly $\alpha\le\|T\|$. For unit vectors $x,y$, self-adjointness gives $\langle Ty,x\rangle=\overline{\langle Tx,y\rangle}$, so $$\begin{aligned}4\operatorname{Re}\langle Tx,y\rangle&=\langle T(x+y),x+y\rangle-\langle T(x-y),x-y\rangle\\&\le\alpha\big(\|x+y\|^2+\|x-y\|^2\big)=4\alpha\end{aligned}$$ by the parallelogram law. If $Tx\ne0$, take $y=Tx/\|Tx\|$: $\|Tx\|\le\alpha$. Hence $\|T\|\le\alpha$.

(3) Let $\lambda=a+ib$. For $b\ne0$: $|\operatorname{Im}\langle(T-\lambda)x,x\rangle|=|b|\,\|x\|^2$, so $\|(T-\lambda)x\|\ge|b|\,\|x\|$. For real $\lambda\gt M$: $\langle(\lambda-T)x,x\rangle\ge(\lambda-M)\|x\|^2$, so $\|(T-\lambda)x\|\ge(\lambda-M)\|x\|$; similarly for $\lambda\lt m$. In each case $T-\lambda$ is bounded below, hence injective with closed range (§8.2), and its range is dense because $\big(\operatorname{ran}(T-\lambda)\big)^\perp=\ker(T-\bar\lambda)=\{0\}$ (§8.3; $\bar\lambda$ satisfies the same bound). So $\lambda\in\rho(T)$. To see $M\in\sigma(T)$, put $A=M-T\ge0$. The form $[x,y]=\langle Ax,y\rangle$ is positive semidefinite, and the proof of Cauchy–Schwarz (§8.3) still gives $|\langle Ax,y\rangle|^2\le\langle Ax,x\rangle\langle Ay,y\rangle$. With $y=Ax$: $$\|Ax\|^4\le\langle Ax,x\rangle\langle A^2x,Ax\rangle\le\langle Ax,x\rangle\|A\|\,\|Ax\|^2,$$ so $\|Ax\|^2\le\|A\|\langle Ax,x\rangle$. Choosing unit $x_n$ with $\langle Tx_n,x_n\rangle\to M$ gives $\langle Ax_n,x_n\rangle\to0$, hence $\|(T-M)x_n\|\to0$: $M$ is an approximate eigenvalue. The same argument with $T-m\ge0$ handles $m$. Finally $r(T)=\max(|m|,|M|)=\|T\|$ by (2).

(4) If $x\in N^\perp$ and $n\in N$, then $\langle Tx,n\rangle=\langle x,Tn\rangle=0$. ∎

📐 Theorem 8.33 — Spectrum of a compact operator (Riesz–Schauder)
Let $X$ be a complex Banach space and $T\in K(X)$.
  1. For $\lambda\ne0$, the eigenspace $\ker(T-\lambda I)$ is finite-dimensional.
  2. For $\lambda\ne0$, $T-\lambda I$ is injective if and only if it is surjective (Fredholm alternative). Hence every nonzero point of $\sigma(T)$ is an eigenvalue.
  3. For every $\varepsilon\gt0$ only finitely many eigenvalues satisfy $|\lambda|\ge\varepsilon$. So $\sigma(T)$ is countable, and $0$ is its only possible accumulation point.
  4. If $\dim X=\infty$, then $0\in\sigma(T)$.

Proof. (1) On $N=\ker(T-\lambda)$, $T=\lambda I$; so the closed unit ball of $N$ is $T(\lambda^{-1}B_N)$, which has compact closure. By Theorem 8.29(5), $\dim N\lt\infty$. (4) is Theorem 8.29(5).

(3) Suppose there were infinitely many distinct eigenvalues $\lambda_1,\lambda_2,\dots$ with $|\lambda_n|\ge\varepsilon$ and eigenvectors $x_n$. Eigenvectors of distinct eigenvalues are linearly independent, so $M_n=\operatorname{span}\{x_1,\dots,x_n\}$ increase strictly. By Riesz's lemma choose unit $y_n\in M_n$ with $\operatorname{dist}(y_n,M_{n-1})\ge\frac12$. Now $(T-\lambda_n)M_n\subseteq M_{n-1}$ (it kills the $x_n$-component) and $TM_m\subseteq M_m$. For $m\lt n$, $$Ty_n-Ty_m=\lambda_ny_n-\big[(\lambda_n-T)y_n+Ty_m\big]$$ with the bracket in $M_{n-1}$, so $\|Ty_n-Ty_m\|\ge|\lambda_n|\cdot\frac12\ge\frac\varepsilon2$. Then $(Ty_n)$ has no convergent subsequence, contradicting compactness.

(2) (Sketch.) One shows first that $\operatorname{ran}(T-\lambda)$ is closed (otherwise there are $x_n$ with $\operatorname{dist}(x_n,\ker(T-\lambda))=1$, bounded, and $(T-\lambda)x_n\to0$; compactness gives $Tx_{n_k}\to y$, then $x_{n_k}\to y/\lambda$, and $y/\lambda$ lies in the kernel, a contradiction). If $T-\lambda$ is injective but not onto, the closed subspaces $\operatorname{ran}(T-\lambda)^k$ decrease strictly, and Riesz's lemma produces a sequence as in (3) contradicting compactness; the converse direction uses the adjoint. For compact self-adjoint operators on a Hilbert space, (2) also follows directly from the spectral theorem (Theorem 8.36(2)). ∎

The Spectral Theorem for Compact Self-Adjoint Operators

The finite-dimensional proof of the spectral theorem finds one eigenvector by maximising $\langle Ax,x\rangle$ on the unit sphere and then restricts $A$ to the orthogonal complement. In infinite dimensions the maximum need not be attained ($M_t$ has none); compactness is exactly what makes it attained.

📐 Theorem 8.34 — A compact self-adjoint operator has an eigenvalue of modulus $\|T\|$
Let $T\ne0$ be a compact self-adjoint operator on a Hilbert space. Then $\|T\|$ or $-\|T\|$ is an eigenvalue of $T$; equivalently, $\sup_{\|x\|=1}|\langle Tx,x\rangle|$ is attained at an eigenvector.

Proof. By Theorem 8.32(2) there are unit vectors $x_n$ with $|\langle Tx_n,x_n\rangle|\to\|T\|$; passing to a subsequence, $\langle Tx_n,x_n\rangle\to\lambda$ with $\lambda=\|T\|$ or $\lambda=-\|T\|$. Since $\lambda$ and $\langle Tx_n,x_n\rangle$ are real, $$\begin{aligned}\|Tx_n-\lambda x_n\|^2&=\|Tx_n\|^2-2\lambda\langle Tx_n,x_n\rangle+\lambda^2\\&\le2\lambda^2-2\lambda\langle Tx_n,x_n\rangle\to0.\end{aligned}$$ By compactness a subsequence has $Tx_{n_k}\to y$. Then $\lambda x_{n_k}=Tx_{n_k}-(Tx_{n_k}-\lambda x_{n_k})\to y$, so $\|y\|=|\lambda|\ne0$, and by continuity $Ty=\lim T(\lambda x_{n_k})=\lambda\lim Tx_{n_k}=\lambda y$. ∎

📐 Theorem 8.35 — Spectral theorem for compact self-adjoint operators (Hilbert–Schmidt)
Let $T$ be a compact self-adjoint operator on a Hilbert space $H$. There is a finite or infinite orthonormal sequence $(e_n)$ of eigenvectors, $Te_n=\lambda_ne_n$, with real $\lambda_n\ne0$ and $|\lambda_1|\ge|\lambda_2|\ge\cdots$, such that:
  1. $Tx=\sum_n\lambda_n\langle x,e_n\rangle e_n$ for every $x\in H$ (norm convergence);
  2. if the sequence is infinite, $\lambda_n\to0$; and $\big\|T-\sum_{n\le N}\lambda_n\langle\cdot,e_n\rangle e_n\big\|=|\lambda_{N+1}|$ (or $0$);
  3. $H=\ker T\oplus\overline{\operatorname{span}}\{e_n\}$; adjoining an orthonormal basis of $\ker T$ gives an orthonormal basis of $H$ consisting of eigenvectors of $T$;
  4. every nonzero eigenvalue of $T$ is one of the $\lambda_n$ and has finite multiplicity (the number of $n$ with $\lambda_n=\lambda$); $\|T\|=|\lambda_1|$; $\ \sigma(T)\setminus\{0\}=\{\lambda_n\}$, and $0\in\sigma(T)$ if $\dim H=\infty$.

Proof. Construction. Put $H_1=H$. Suppose $e_1,\dots,e_{n-1}$ are constructed and let $H_n=\{e_1,\dots,e_{n-1}\}^\perp$. Each $e_k$ spans a $T$-invariant line, so $H_n$ is $T$-invariant (Theorem 8.32(4)), and $T_n=T|_{H_n}$ is compact and self-adjoint on the Hilbert space $H_n$. If $T_n=0$, stop. Otherwise Theorem 8.34 gives a unit $e_n\in H_n$ with $Te_n=\lambda_ne_n$, $|\lambda_n|=\|T_n\|$. Since $H_{n+1}\subseteq H_n$, $|\lambda_{n+1}|=\|T_{n+1}\|\le\|T_n\|=|\lambda_n|$.

$\lambda_n\to0$. If $|\lambda_n|\ge\varepsilon\gt0$ for infinitely many (hence, by monotonicity, all) $n$, then $\|Te_n-Te_m\|^2=\lambda_n^2+\lambda_m^2\ge2\varepsilon^2$, and $(Te_n)$ has no convergent subsequence — impossible for compact $T$.

Expansion. For $x\in H$ let $y_N=x-\sum_{n\le N}\langle x,e_n\rangle e_n$. Then $y_N\in H_{N+1}$ and $\|y_N\|\le\|x\|$ (Bessel's inequality, §8.3), so $$\begin{aligned}\Big\|Tx-\sum_{n\le N}\lambda_n\langle x,e_n\rangle e_n\Big\|&=\|Ty_N\|\\&\le|\lambda_{N+1}|\,\|x\|\to0.\end{aligned}$$ This proves (1) and the upper bound in (2); equality holds at $x=e_{N+1}$. If the process stopped at step $N+1$, $Ty_N=0$ and the sum is finite.

Decomposition. If $x\perp e_n$ for all $n$, then $Tx=0$ by (1). Conversely, if $Tx=0$, then $\lambda_n\langle x,e_n\rangle=\langle x,Te_n\rangle=\langle Tx,e_n\rangle=0$, so $x\perp e_n$. Thus $\ker T=\{e_n\}^\perp$ and (3) follows from the projection theorem (§8.3).

Eigenvalues. If $Tx=\mu x$ with $\mu\ne0$, $x\ne0$, then $\mu\langle x,e_n\rangle=\langle Tx,e_n\rangle=\lambda_n\langle x,e_n\rangle$, so $\langle x,e_n\rangle=0$ unless $\lambda_n=\mu$; as $x=\mu^{-1}Tx\in\overline{\operatorname{span}}\{e_n\}$, $x$ is a combination of the finitely many $e_n$ with $\lambda_n=\mu$ (finitely many because $\lambda_n\to0$). So $\mu$ is some $\lambda_n$ and its eigenspace is $\operatorname{span}\{e_n:\lambda_n=\mu\}$. $\|T\|=\|T_1\|=|\lambda_1|$. A nonzero $\mu$ outside $\{\lambda_n\}$ is in $\rho(T)$ by Theorem 8.36(2) below, and $0\in\sigma(T)$ in infinite dimensions by Theorem 8.29(5). ∎

In words: a compact self-adjoint operator is a diagonal operator $D_\lambda$ with $\lambda_n\to0$, written in a suitable orthonormal basis. Interactive 1 runs the proof as an algorithm.

Interactive 1 — Finding Eigenvalues the Way the Proof Does: Power Iteration and Deflation

The integral operator $Tu(x)=\int_0^1K(x,t)u(t)\,dt$ (discretised on 160 points) is applied again and again to a starting function: $v\mapsto Tv/\|Tv\|$. The iterate (red) turns into the eigenfunction of largest $|\lambda|$ (dashed) and the Rayleigh quotient $\langle Tv,v\rangle$ climbs to $\lambda_1$ — Theorem 8.34 in action. Then the found eigenvector is projected out (deflation: restrict $T$ to $\{e_1\}^\perp$, as in the proof of Theorem 8.35) and the search starts again for $\lambda_2$, then $\lambda_3$. Bottom: the residual $\|Tv-\langle Tv,v\rangle v\|$ on a log scale; it falls like $|\lambda_{j+1}/\lambda_j|^k$.

searching for—
iteration k—
Rayleigh ⟨Tv, v⟩—
exact λ—
📐 Theorem 8.36 — Consequences: Rayleigh–Ritz, the Fredholm alternative, Hilbert–Schmidt norm
Let $T$ be compact and self-adjoint with eigen-data $(\lambda_n,e_n)$ as in Theorem 8.35, and let $P_0$ be the orthogonal projection onto $\ker T$.
  1. Rayleigh–Ritz. If $\lambda_1^+\ge\lambda_2^+\ge\cdots\gt0$ are the positive eigenvalues (with multiplicity) and $e_1^+,e_2^+,\dots$ their eigenvectors, then $$\lambda_k^+=\max_{\substack{\|x\|=1\\ x\perp e_1^+,\dots,e_{k-1}^+}}\langle Tx,x\rangle.$$ In particular $\langle Tx,x\rangle\le\lambda_1^+\|x\|^2$, and any trial vector gives a lower bound for $\lambda_1^+$.
  2. Fredholm alternative. Let $\mu\ne0$. If $\mu$ is not an eigenvalue, then $(T-\mu)u=f$ has, for every $f$, exactly one solution $$u=-\frac f\mu+\sum_n\frac{\lambda_n\langle f,e_n\rangle}{\mu(\lambda_n-\mu)}e_n,$$ and $\|(T-\mu)^{-1}\|=1/\operatorname{dist}(\mu,\sigma(T))$. If $\mu$ is an eigenvalue, a solution exists if and only if $f\perp\ker(T-\mu)$, and then the solutions form a translate of $\ker(T-\mu)$.
  3. Hilbert–Schmidt. $T$ is Hilbert–Schmidt iff $\sum_n\lambda_n^2\lt\infty$, and $\|T\|_{HS}^2=\sum_n\lambda_n^2$ (with multiplicity). For an integral operator with Hermitian kernel $K\in L^2$: $\sum_n\lambda_n^2=\int\!\!\int|K|^2$.
  4. Mercer's theorem (statement). If $K$ is continuous on $[a,b]^2$, Hermitian, and $T_K\ge0$, then $K(x,t)=\sum_n\lambda_ne_n(x)\overline{e_n(t)}$ absolutely and uniformly, and $\sum_n\lambda_n=\int_a^bK(x,x)\,dx$.

Proof. (1) Write $x=x_0+\sum_nc_ne_n$ with $x_0\in\ker T$. Then $\langle Tx,x\rangle=\sum_n\lambda_n|c_n|^2$. If $x\perp e_1^+,\dots,e_{k-1}^+$, every term with $\lambda_n\gt0$ has $\lambda_n\le\lambda_k^+$ and the terms with $\lambda_n\lt0$ are $\le0$, so $\langle Tx,x\rangle\le\lambda_k^+\sum|c_n|^2\le\lambda_k^+\|x\|^2$; equality at $x=e_k^+$.

(2) Expand $f=P_0f+\sum\langle f,e_n\rangle e_n$ and look for $u=u_0+\sum c_ne_n$. Then $(T-\mu)u=-\mu u_0+\sum(\lambda_n-\mu)c_ne_n$, so the equation splits into $-\mu u_0=P_0f$ and $(\lambda_n-\mu)c_n=\langle f,e_n\rangle$. If $\mu\ne\lambda_n$ for all $n$, then $|\lambda_n-\mu|\ge\delta\gt0$ (because $\lambda_n\to0\ne\mu$), $c_n=\langle f,e_n\rangle/(\lambda_n-\mu)$ is square-summable, and $$u=-\frac{P_0f}\mu+\sum_n\frac{\langle f,e_n\rangle}{\lambda_n-\mu}e_n;$$ substituting $P_0f=f-\sum\langle f,e_n\rangle e_n$ and $\frac1{\lambda_n-\mu}+\frac1\mu=\frac{\lambda_n}{\mu(\lambda_n-\mu)}$ gives the formula. The inverse is "diagonal" with entries $-1/\mu$ and $1/(\lambda_n-\mu)$, whose supremum is $1/\operatorname{dist}(\mu,\sigma(T))$. If $\mu=\lambda_k$ for some $k$, the equations with $\lambda_n=\mu$ read $0=\langle f,e_n\rangle$ — the condition $f\perp\ker(T-\mu)$ — and leave those $c_n$ free.

(3) Compute $\|T\|_{HS}$ in the orthonormal basis of eigenvectors of Theorem 8.35(3): $\sum\|Te_n\|^2=\sum\lambda_n^2$; for integral operators combine with Theorem 8.30(4). (4) is quoted without proof. ∎

✏️ Example 3 — A degenerate kernel: $K(x,t)=x+t$ on $[0,1]$
Let $Tu(x)=\int_0^1(x+t)u(t)\,dt$ on $L^2[0,1]$ (real scalars). Find all eigenvalues, normalised eigenfunctions, $\ker T$, $\sigma(T)$ and $\|T\|$. Is $T$ positive?
Structure $Tu=x\langle u,1\rangle+\langle u,t\rangle$, so $\operatorname{ran}T\subseteq\operatorname{span}\{1,x\}$: $T$ has rank $2$ (compact), and its real symmetric kernel makes it self-adjoint. An eigenfunction with $\lambda\ne0$ lies in $\operatorname{ran}T$, so $u=a+bx$.
Reduce to a matrix $T(a+bx)=\big(\tfrac a2+\tfrac b3\big)+\big(a+\tfrac b2\big)x$. In coordinates $(a,b)$ this is $A=\begin{pmatrix}\frac12&\frac13\\1&\frac12\end{pmatrix}$, with trace $1$ and determinant $\frac14-\frac13=-\frac1{12}$. So $\lambda^2-\lambda-\frac1{12}=0$: $$\lambda_\pm=\tfrac12\pm\tfrac1{\sqrt3}\approx1.0774,\ -0.0774.$$
Eigenfunctions $(\tfrac12-\lambda_\pm)a+\tfrac b3=0$ gives $b=\pm\sqrt3\,a$: $u_\pm=1\pm\sqrt3x$, with $\|u_\pm\|^2=2\pm\sqrt3$. Normalised: $e_\pm=(1\pm\sqrt3x)/\sqrt{2\pm\sqrt3}$. They are orthogonal, as Theorem 8.32(1) predicts: $\int_0^1(1-3x^2)dx=0$.
Kernel and spectrum $\ker T=\{u:\int u=\int tu=0\}=\{1,t\}^\perp$, infinite-dimensional. $\sigma(T)=\{0,\ \tfrac12+\tfrac1{\sqrt3},\ \tfrac12-\tfrac1{\sqrt3}\}$ and $\|T\|=\tfrac12+\tfrac1{\sqrt3}\approx1.0774$ (checked numerically). $Tu=\lambda_+\langle u,e_+\rangle e_++\lambda_-\langle u,e_-\rangle e_-$.
Not positive $\lambda_-\lt0$; concretely $\langle Tu,u\rangle=2\langle u,1\rangle\langle u,t\rangle$, which for $u=3-5t$ equals $2\cdot\frac12\cdot(-\frac16)=-\frac16$. Checks: $\lambda_++\lambda_-=1=\int_0^1K(x,x)\,dx$ and $\lambda_+^2+\lambda_-^2=\frac76=\int\!\!\int(x+t)^2$.
✏️ Example 4 — The kernel $\min(x,t)$: eigenvalues from an ODE
Let $(Tu)(x)=\int_0^1\min(x,t)\,u(t)\,dt$ on $L^2[0,1]$. (a) Show $T$ is compact, self-adjoint, positive and injective. (b) Find its eigenvalues and eigenfunctions. (c) Deduce that $\{\sqrt2\sin((n-\tfrac12)\pi x)\}_{n\ge1}$ is an orthonormal basis of $L^2[0,1]$, find $\|T\|$, and compare with the Rayleigh quotient of $u=x$. (d) Use Mercer's theorem to sum $\sum_{n\ge1}(2n-1)^{-2}$.
(a) The kernel is continuous, real and symmetric: $T$ is Hilbert–Schmidt (compact) and self-adjoint. Since $\min(x,t)=\int_0^1\mathbf 1_{s\lt x}\mathbf 1_{s\lt t}\,ds$, Fubini gives $\langle Tu,u\rangle=\int_0^1\Big(\int_s^1u\Big)^2ds\ge0$. Write $v=Tu=\int_0^xtu(t)\,dt+x\int_x^1u(t)\,dt$. Differentiating, $v'(x)=\int_x^1u$, so $$v(0)=0,\quad v'(1)=0,\quad v''=-u.$$ Hence $Tu=0$ forces $u=-v''=0$: $T$ is injective.
(b) ODE $Tu=\lambda u$ with $\lambda\ne0$ makes $u=v/\lambda$ twice differentiable with $\lambda u''=-u$, $u(0)=0$, $u'(1)=0$. As $T\ge0$, $\lambda\gt0$; write $\lambda=1/\omega^2$. Then $u=\sin\omega x$ and $u'(1)=\omega\cos\omega=0$, so $\omega=(n-\frac12)\pi$: $$\lambda_n=\frac1{(n-\frac12)^2\pi^2}=\frac4{(2n-1)^2\pi^2},$$ $$e_n=\sqrt2\sin\big((n-\tfrac12)\pi x\big).$$
(c) $\ker T=\{0\}$, so by Theorem 8.35(3) the $e_n$ form an orthonormal basis. $\|T\|=\lambda_1=\frac4{\pi^2}\approx0.4053$. Rayleigh: $\langle Tx,x\rangle/\|x\|^2=\frac{2/15}{1/3}=\frac25=0.4\le0.4053$ — a lower bound accurate to about $1.3\%$ (Theorem 8.36(1)); the trial $u=x(2-x)$, which satisfies both boundary conditions, gives $\frac{17}{42}\approx0.40476$.
(d) Mercer $T\ge0$ with continuous kernel: $\sum_n\lambda_n=\int_0^1\min(x,x)\,dx=\frac12$, i.e. $\frac4{\pi^2}\sum\frac1{(2n-1)^2}=\frac12$, so $\sum\frac1{(2n-1)^2}=\frac{\pi^2}8$. Likewise $\sum\lambda_n^2=\int\!\!\int\min(x,t)^2=\frac16$ gives $\sum\frac1{(2n-1)^4}=\frac{\pi^4}{96}$.
Interactive 2 — A Kernel Rebuilt from Its Eigenfunctions

Left: the kernel $K(x,t)$ on $[0,1]^2$ (blue positive, red negative) and its rank-$N$ truncation $K_N(x,t)=\sum_{n\le N}\lambda_ne_n(x)e_n(t)$, the kernel of $T_N=\sum_{n\le N}\lambda_n\langle\cdot,e_n\rangle e_n$. Bottom left: the eigenvalues, those used in blue. Right: the slice $t=0.7$ of $K$ (black) and $K_N$ (red). The operator-norm error is exactly $|\lambda_{N+1}|$ (Theorem 8.35(2)) and the Hilbert–Schmidt error is $\big(\sum_{n\gt N}\lambda_n^2\big)^{1/2}$; for a positive continuous kernel the uniform error also tends to $0$ (Mercer).

N—
‖T − T_N‖ = |λ_(N+1)|—
‖T − T_N‖_HS—
max |K − K_N|—
🧭 Problem-solving strategy — eigenvalues and norms of integral operators
  1. Degenerate kernel $K(x,t)=\sum_{i=1}^ra_i(x)b_i(t)$: the range lies in $\operatorname{span}\{a_i\}$. Put $u=\sum c_ia_i$; then $Tu=\lambda u$ becomes the $r\times r$ eigenproblem $Ac=\lambda c$ with $A_{ij}=\int b_ia_j$. All other "eigenvalues" are $0$, with infinite-dimensional kernel.
  2. Self-adjoint? Real symmetric (or Hermitian) kernel ⇒ real eigenvalues, orthogonal eigenfunctions, $\|T\|=\max|\lambda|$. Without symmetry, $\|T\|$ can exceed every $|\lambda|$ (the Volterra operator).
  3. Green's-function kernels ($\min(x,t)$, $\min(x,t)-xt$, …): differentiate $v=Tu$ twice to get a boundary value problem; the eigenvalues of $T$ are the reciprocals of the Sturm–Liouville eigenvalues.
  4. Checks: $\sum\lambda_n=\int K(x,x)\,dx$ (finite rank, or Mercer for positive continuous kernels) and $\sum\lambda_n^2=\int\!\!\int|K|^2$. A trial function $u$ gives $\lambda_1^+\ge\langle Tu,u\rangle/\|u\|^2$.
  5. Solvability of $u-\lambda Tu=f$: unique iff $1/\lambda$ is not an eigenvalue of $T$; at a characteristic value, solvable iff $f$ is orthogonal to the eigenspace.

Applications: Integral Equations and Sturm–Liouville Problems

✏️ Example 5 — A Fredholm integral equation of the second kind
Consider $u(x)=f(x)+\lambda\int_0^1(x+t)\,u(t)\,dt$ on $L^2[0,1]$. (a) For which real $\lambda$ does it have exactly one solution for every $f$? (b) Solve it for $f\equiv1$ and $\lambda=\frac12$. (c) At $\lambda^*=4\sqrt3-6$, for which $f$ is there a solution? Is there one for $f\equiv1$?
(a) With $T$ as in Example 3 the equation is $(I-\lambda T)u=f$; for $\lambda\ne0$ it is $(T-\frac1\lambda)u=-\frac f\lambda$. By Theorem 8.36(2) it is uniquely solvable for all $f$ iff $\frac1\lambda\ne\lambda_\pm=\frac12\pm\frac1{\sqrt3}$, i.e. $$\lambda\ne\frac1{\frac12\pm\frac1{\sqrt3}}=-6\pm4\sqrt3\approx0.9282,\ -12.928.$$ ($\lambda=0$ is trivially fine.) Equivalently $\det(I-\lambda A)=1-\lambda-\frac{\lambda^2}{12}\ne0$ with $A$ from Example 3.
(b) Degenerate-kernel method $u=1+\lambda(\alpha x+\beta)$ with $\alpha=\int_0^1u$, $\beta=\int_0^1tu$. Integrating against $1$ and $t$ with $\lambda=\frac12$: $\alpha=1+\frac\alpha4+\frac\beta2$ and $\beta=\frac12+\frac\alpha6+\frac\beta4$. From the first, $\beta=\frac32\alpha-2$; substituting, $\frac{23}{24}\alpha=2$, so $\alpha=\frac{48}{23}$, $\beta=\frac{26}{23}$, and $$u(x)=\frac{36+24x}{23}.$$
(b) Check $\int_0^1u=\frac{36+12}{23}=\frac{48}{23}=\alpha$ ✓. The spectral formula $u=f+\sum_\pm\frac{\lambda\lambda_\pm\langle f,e_\pm\rangle}{1-\lambda\lambda_\pm}e_\pm$ (Theorem 8.36(2) rewritten for $u-\lambda Tu=f$) gives the same function (verified numerically: $u(0)\approx1.5652$, $u(1)\approx2.6087$).
(c) At $\lambda^*$, $\frac1{\lambda^*}=\lambda_+$ is an eigenvalue with eigenspace $\operatorname{span}\{1+\sqrt3x\}$. A solution exists iff $f\perp e_+$, i.e. $\int_0^1f(t)(1+\sqrt3t)\,dt=0$, and then solutions differ by multiples of $1+\sqrt3x$. For $f\equiv1$ the integral is $1+\frac{\sqrt3}2\ne0$: no solution.

The most important application turns a differential operator, which is unbounded, into a compact self-adjoint integral operator by inverting it.

📐 Theorem 8.37 — Regular Sturm–Liouville problems
Let $p\in C^1[a,b]$ with $p\gt0$, $q\in C[a,b]$ real, and $Lu=-(pu')'+qu$ with separated boundary conditions $$\alpha_1u(a)+\alpha_2u'(a)=0,\qquad\beta_1u(b)+\beta_2u'(b)=0$$ ($(\alpha_1,\alpha_2)\ne0\ne(\beta_1,\beta_2)$, real). Assume $0$ is not an eigenvalue of $L$ with these conditions. Then:
  1. there is a continuous, real, symmetric Green's function $G$ such that for $f\in C[a,b]$ the unique solution of $Lu=f$ with the boundary conditions is $u(x)=\int_a^bG(x,t)f(t)\,dt$; explicitly $G(x,t)=u_a(\min(x,t))\,u_b(\max(x,t))/C$, where $u_a$, $u_b$ solve $Lu=0$ with the condition at $a$, resp. $b$, and $C=p\,(u_a'u_b-u_au_b')$ is constant;
  2. $T_G$ is compact, self-adjoint and injective on $L^2[a,b]$;
  3. the eigenvalues of the Sturm–Liouville problem $Lu=\lambda u$ are real and simple, form a sequence $\lambda_1\lt\lambda_2\lt\cdots$ with $|\lambda_n|\to\infty$, and the normalised eigenfunctions form an orthonormal basis of $L^2[a,b]$.

Proof (sketch). (1) If $u_a,u_b$ were dependent, $u_a$ would satisfy both conditions and $0$ would be an eigenvalue; so they are independent and $C\ne0$; $C$ is constant because $(p(u_a'u_b-u_au_b'))'$ $=u_b(pu_a')'-u_a(pu_b')'$ $=u_bqu_a-u_aqu_b=0$. Direct differentiation of $u(x)=\frac1C\big(u_b(x)\int_a^xu_af+u_a(x)\int_x^bu_bf\big)$ — exactly as in Example 4 — gives $Lu=f$ and the boundary conditions; uniqueness holds because $0$ is not an eigenvalue. $G$ is visibly symmetric. (2) A continuous symmetric kernel gives a compact (Hilbert–Schmidt) self-adjoint operator. If $T_Gf=0$ for continuous $f$, then $f=L(T_Gf)=0$; for $f\in L^2$ the same identity holds a.e. (with $pu'$ absolutely continuous), so $T_G$ is injective. (3) For $\lambda\ne0$: $Lu=\lambda u$ with the boundary conditions iff $u=T_G(\lambda u)$ iff $T_Gu=\frac1\lambda u$. So the eigenvalues of $L$ are the reciprocals $1/\mu_n$ of the nonzero eigenvalues of $T_G$ (real), and since $\ker T_G=\{0\}$, Theorem 8.35(3) makes the eigenfunctions an orthonormal basis; $\mu_n\to0$ gives $|\lambda_n|\to\infty$. Simplicity: two eigenfunctions for the same $\lambda$ both satisfy the condition at $a$, so their Wronskian vanishes at $a$ and they are dependent. That the $\lambda_n$ are bounded below (so $\lambda_n\to+\infty$) follows from $\langle Lu,u\rangle=\int(p|u'|^2+q|u|^2)+$ boundary terms; we omit it. ∎

✏️ Example 6 — The vibrating string: $-u''=\lambda u$, $u(0)=u(1)=0$
(a) Find the Green's function of $-u''$ with $u(0)=u(1)=0$. (b) Using Theorem 8.37, show that $\{\sqrt2\sin n\pi x\}_{n\ge1}$ is an orthonormal basis of $L^2[0,1]$. (c) Use Mercer's theorem and the Hilbert–Schmidt identity to evaluate $\sum1/n^2$ and $\sum1/n^4$.
(a) $p=1$, $q=0$; $u_a=x$ (vanishes at $0$), $u_b=1-x$ (vanishes at $1$), $C=u_a'u_b-u_au_b'=(1-x)+x=1$. So $G(x,t)=\min(x,t)\big(1-\max(x,t)\big)$ $=\min(x,t)-xt$. Check: $u=\int_0^1Gf$ is $(1-x)\int_0^xtf+x\int_x^1(1-t)f$, whose derivative is $-\int_0^xtf+\int_x^1(1-t)f$ and second derivative $-xf(x)-(1-x)f(x)=-f(x)$.
(b) $0$ is not an eigenvalue ($u''=0$, $u(0)=u(1)=0$ gives $u=0$). The eigenvalues of $-u''=\lambda u$ with these conditions: for $\lambda=\omega^2\gt0$, $u=\sin\omega x$ with $\sin\omega=0$, so $\lambda_n=n^2\pi^2$; $\lambda\le0$ gives only $u=0$. By Theorem 8.37 the normalised eigenfunctions $\sqrt2\sin n\pi x$ form an orthonormal basis — a second proof of the completeness stated in §8.3, now from the spectral theorem. $T_G$ has eigenvalues $\mu_n=\frac1{n^2\pi^2}$ and $\|T_G\|=\frac1{\pi^2}\approx0.1013$.
(c) $T_G\ge0$ (all $\mu_n\gt0$) with continuous kernel, so Mercer gives $\sum\mu_n=\int_0^1G(x,x)\,dx=\int_0^1x(1-x)\,dx=\frac16$, i.e. $\sum\frac1{n^2}=\frac{\pi^2}6$. And $\sum\mu_n^2=\int\!\!\int G^2=\frac1{90}$ (computed symbolically) gives $\sum\frac1{n^4}=\frac{\pi^4}{90}$ — the values found with Parseval in §8.3.
✏️ Example 7 — True or false? Prove or give a counterexample
(a) Every compact operator on an infinite-dimensional Hilbert space has an eigenvalue. (b) Every bounded self-adjoint operator has an eigenvalue. (c) A compact self-adjoint operator with $\sigma(T)=\{0\}$ is the zero operator. (d) If $T$ is compact, self-adjoint and injective on an infinite-dimensional Hilbert space $H$, then $0\in\sigma(T)$, $0$ is not an eigenvalue, and $H$ is separable. (e) A compact operator with $\lim\|T^n\|^{1/n}=0$ is the zero operator.
(a) False The Volterra operator is compact with $\sigma_p(V)=\varnothing$ (Example 2(d)). It is not self-adjoint, so Theorem 8.34 does not apply.
(b) False $M_t$ on $L^2[0,1]$ is self-adjoint with $\sigma=[0,1]$ and no eigenvalues (Example 2(b)); it is not compact.
(c) True $\|T\|=r(T)=0$ by Theorem 8.32(3). (Self-adjointness is essential: compare (e).)
(d) True $0\in\sigma(T)$ by Theorem 8.29(5); $0$ is not an eigenvalue by injectivity. By Theorem 8.35(3) with $\ker T=\{0\}$, the eigenvectors $(e_n)$ — a countable family — form an orthonormal basis, so $H$ is separable (§8.3).
(e) False $V$ again: $\|V^n\|^{1/n}\le(1/n!)^{1/n}\to0$, yet $\|V\|=\frac2\pi$. Quasinilpotent nonzero operators exist only among non-normal operators.
⚠️ Counterexamples and pitfalls
  1. Bounded is not compact. The identity and every isometry of an infinite-dimensional space, the shifts, $M_t$ on $L^2[0,1]$, and projections onto infinite-dimensional subspaces are bounded but not compact.
  2. Strong limits of finite-rank operators need not be compact: the truncations $P_N\to I$ pointwise on $\ell^2$, but $\|P_N-I\|=1$. Only norm limits preserve compactness.
  3. $0$ is always in the spectrum of a compact operator in infinite dimensions, but it need not be an eigenvalue (diagonal $(\frac1n)$), and when it is, its eigenspace may be infinite-dimensional (kernel $x+t$). Nonzero eigenvalues have finite multiplicity.
  4. A compact operator may have no eigenvalues (Volterra, $\sigma=\{0\}$); a self-adjoint operator may have no eigenvalues (multiplication by $t$, $\sigma=[0,1]$). The spectral theorem needs both compactness and self-adjointness (or normality).
  5. $\|T\|=r(T)$ only for self-adjoint (normal) operators. $\begin{pmatrix}0&1\\0&0\end{pmatrix}$ has $r=0$, $\|\cdot\|=1$; $V$ has $r=0$, $\|V\|=\frac2\pi$. For a non-symmetric kernel, do not compute $\|T\|$ as $\max|\lambda|$.
  6. Real spectrum does not imply self-adjoint (nilpotent matrices, $V$).
  7. The spectrum is more than the eigenvalues in infinite dimensions: the right shift has $\sigma(S)$ = closed unit disc but no eigenvalues; truncating it to $N\times N$ matrices gives nilpotent matrices with spectrum $\{0\}$ — finite sections can mislead badly for non-self-adjoint operators.
  8. Eigenfunction expansions converge in $L^2$, not necessarily pointwise; uniform convergence of $\sum\lambda_ne_n(x)\overline{e_n(t)}$ needs Mercer's hypotheses (continuous kernel, $T\ge0$).
  9. Characteristic values versus eigenvalues: $u=f+\lambda Tu$ fails to be uniquely solvable when $\lambda=1/\mu$ for an eigenvalue $\mu\ne0$ of $T$ — not when $\lambda$ itself is an eigenvalue. A classic exam slip.

Exercises

I
Compact operators.
  • 1 Which diagonal operators $D_\lambda$ on $\ell^2$ are compact? (a) $\lambda_n=n^{-1/2}$; (b) $\lambda_n=(-1)^n$; (c) $\lambda_n=2^{-n}$; (d) $\lambda_n=1$ if $n$ is prime and $0$ otherwise.
  • 2 Prove that the weighted shift $Tx=(0,x_1,\tfrac{x_2}2,\tfrac{x_3}3,\dots)$ is compact on $\ell^2$.
  • 3 Let $\varphi\in C[0,1]$. Prove that $M_\varphi$ is compact on $L^2[0,1]$ only if $\varphi\equiv0$.
  • 4 Prove the ideal property: if $T$ is compact and $S$ bounded, $ST$ and $TS$ are compact. Deduce that no compact operator on $\ell^2$ is invertible in $B(\ell^2)$.
  • 5 Let $T\in B(H)$. Prove that if $T^*T$ is compact then $T$ is compact.
  • 6 Compute $\|V\|_{HS}$ for the Volterra operator and deduce an upper bound for $\|V\|$. Compare with $\|V\|=\frac2\pi$.
  • 7 Give a sequence of finite-rank operators converging pointwise (but not in norm) to a non-compact operator.
  • 8 Let $K\in C([0,1]^2)$. Prove that $(Tf)(x)=\int_0^1K(x,t)f(t)\,dt$ is compact on $(C[0,1],\|\cdot\|_\infty)$.
1 By Theorem 8.30(5), compact iff $\lambda_n\to0$: (a) yes (not Hilbert–Schmidt, since $\sum\frac1n=\infty$ — compact operators need not be Hilbert–Schmidt); (b) no; (c) yes; (d) no — there are infinitely many primes, so $\lambda_n=1$ infinitely often.
2 Let $T_N$ use the weights $1,\frac12,\dots,\frac1N$ and then $0$: finite rank, and $\|T-T_N\|=\sup_{n\gt N}\frac1n=\frac1{N+1}\to0$. By Theorem 8.29(2), $T$ is compact. (It is also Hilbert–Schmidt: $\sum\|Te_n\|^2=\sum\frac1{n^2}$.)
3 If $\varphi(t_0)\ne0$, continuity gives an interval $J$ with $|\varphi|\ge c\gt0$ on $J$. Take an orthonormal sequence $(u_n)$ of functions supported in $J$ (e.g. rescaled sines). Then $\|M_\varphi u_n\|\ge c$, so $M_\varphi u_n\not\to0$, contradicting Theorem 8.30(1).
4 See Theorem 8.29(3). If $T$ compact had a bounded inverse, $I=T^{-1}T$ would be compact on the infinite-dimensional $\ell^2$, contradicting Theorem 8.29(5).
5 Let $\|x_n\|\le1$; pass to a subsequence with $T^*Tx_n$ convergent. Then $$\begin{aligned}\|Tx_n-Tx_m\|^2&=\langle T^*T(x_n-x_m),x_n-x_m\rangle\\&\le2\|T^*T(x_n-x_m)\|\to0,\end{aligned}$$ so $(Tx_n)$ is Cauchy and converges.
6 $\|V\|_{HS}^2=\int_0^1\!\int_0^x1\,dt\,dx=\frac12$, so $\|V\|\le\frac1{\sqrt2}\approx0.7071$; the true value $\frac2\pi\approx0.6366$ (computed numerically from the singular values) is smaller, as it must be.
7 The coordinate projections $P_N$ on $\ell^2$: $P_Nx\to x$ for every $x$, but $\|P_N-I\|=\|e_{N+1}\|=1$, and the limit $I$ is not compact.
8 For $\|f\|_\infty\le1$: $|Tf(x)|\le\max|K|$ (uniformly bounded), and $|Tf(x)-Tf(y)|\le\max_t|K(x,t)-K(y,t)|$, which is small when $|x-y|$ is small by uniform continuity of $K$ (equicontinuous). By the Arzelà–Ascoli theorem (§4.2), $T(B)$ has compact closure.
II
Spectra.
  • 1 For $A=\begin{pmatrix}0&1\\0&0\end{pmatrix}$ and $B=\begin{pmatrix}2&1\\1&2\end{pmatrix}$ on $\mathbb{C}^2$ (Euclidean norm), find the spectrum, the norm and the spectral radius.
  • 2 Find $\sigma(D)$ and the eigenvalues of the diagonal operator $D$ on $\ell^2$ with $\lambda_n=1+\frac{(-1)^n}n$.
  • 3 Prove that $\sigma(T^*)=\{\bar\lambda:\lambda\in\sigma(T)\}$ for $T\in B(H)$.
  • 4 Find the spectrum and eigenvalues of $M_\varphi$ on $L^2[0,1]$ for (a) $\varphi(t)=t^2$; (b) $\varphi=\mathbf 1_{[0,1/2]}$.
  • 5 Show that $V^n$ has kernel $\frac{(x-t)^{n-1}}{(n-1)!}$ for $t\le x$ (and $0$ otherwise), that $\|V^n\|\le\frac1{n!}$, and hence that $\sigma(V)=\{0\}$.
  • 6 Prove that the spectrum of a unitary operator lies on the unit circle.
  • 7 Let $T\ge0$ be self-adjoint. Prove that $\sigma(T)\subseteq[0,\|T\|]$ and $\|T\|\in\sigma(T)$.
  • 8 On $\ell^2$ let $Tx=(x_2,x_1,x_4,x_3,\dots)$. Show $T$ is self-adjoint and unitary, find $\sigma(T)$ with multiplicities, and decide whether $T$ is compact.
1 $A$: $\sigma=\{0\}$, $r(A)=0$, $\|A\|=1$ ($Ae_2=e_1$). $B$: symmetric with eigenvalues $1,3$, so $\|B\|=r(B)=3$ — equality because $B$ is self-adjoint.
2 $\sigma_p=\{\lambda_n\}=\{0,\frac32,\frac23,\frac54,\frac45,\dots\}$ and $\sigma=\{\lambda_n\}\cup\{1\}$ (Example 2(a)). The limit point $1$ is in the spectrum but is not an eigenvalue ($\lambda_n\ne1$ for all $n$).
3 $(T-\lambda)^*=T^*-\bar\lambda$, and $S$ is invertible in $B(H)$ iff $S^*$ is: $(S^{-1})^*S^*=(SS^{-1})^*=I$ and similarly on the other side, and $S^{**}=S$.
4 (a) $\sigma=$ range of $t^2$ $=[0,1]$, no eigenvalues (as for $M_t$). (b) $\varphi$ takes the values $0,1$: $\sigma=\{0,1\}$, both eigenvalues, with infinite-dimensional eigenspaces $L^2(\tfrac12,1]$ and $L^2[0,\tfrac12]$; for $\mu\notin\{0,1\}$, $M_{1/(\varphi-\mu)}$ is a bounded inverse.
5 Induction: $V^{n+1}u(x)=\int_0^x\!\int_0^s\frac{(s-t)^{n-1}}{(n-1)!}u(t)\,dt\,ds$, which equals $\int_0^x\frac{(x-t)^n}{n!}u(t)\,dt$. The Hilbert–Schmidt norm is $\|V^n\|_{HS}^2=\frac1{((n-1)!)^2(2n-1)2n}$, so $\|V^n\|\le\frac1{(n-1)!\sqrt{2n(2n-1)}}\le\frac1{n!}$ (as $2n(2n-1)\ge n^2$). Stirling: $(n!)^{-1/n}\to0$, so $r(V)=0$; $0\in\sigma(V)$ by compactness.
6 $\|U\|=\|U^{-1}\|=1$, so $\sigma(U)$ and $\sigma(U^{-1})$ lie in the closed unit disc. For $\lambda\ne0$, $U-\lambda=-\lambda U(U^{-1}-\lambda^{-1})$, so $\lambda\in\sigma(U)$ iff $\lambda^{-1}\in\sigma(U^{-1})$; hence $|\lambda|\le1$ and $|\lambda^{-1}|\le1$. ($0\notin\sigma(U)$.)
7 $m\ge0$ and $M=\|T\|$ (Theorem 8.32(2)), so $\sigma(T)\subseteq[m,M]\subseteq[0,\|T\|]$ and $M=\|T\|\in\sigma(T)$ by Theorem 8.32(3).
8 $\langle Tx,y\rangle=\sum(x_2\bar y_1+x_1\bar y_2+\cdots)=\langle x,Ty\rangle$ and $T^2=I$, so $T=T^*=T^{-1}$. $T(e_{2k-1}\pm e_{2k})=\pm(e_{2k-1}\pm e_{2k})$: eigenvalues $\pm1$, each of infinite multiplicity, and $\sigma(T)=\{-1,1\}$ (e.g. since $T^2=I$, $(T-\mu)(T+\mu)=(1-\mu^2)I$). Not compact: nonzero eigenvalues of compact operators have finite multiplicity.
III
The spectral theorem in practice.
  • 1 On $L^2[-1,1]$ let $K(x,t)=1+3xt$. Find the eigenvalues with multiplicities and show $T=2P$ for an orthogonal projection $P$.
  • 2 On $L^2[0,2\pi]$ let $K(x,t)=\cos(x-t)$. Find the nonzero eigenvalues, their multiplicities and $\|T\|$.
  • 3 On $L^2[-1,1]$ let $K(x,t)=xt+x^2t^2$. Find the nonzero eigenvalues and $\|T\|$.
  • 4 For $K=\min(x,t)$ on $[0,1]$ (Example 4), find the smallest $N$ such that the $N$-term truncation $T_N$ satisfies $\|T-T_N\|\lt0.01$.
  • 5 Prove that $\langle Tu,u\rangle=\int_0^1\big(\int_s^1u\big)^2ds$ for $K=\min(x,t)$ by writing $\min(x,t)$ as an integral.
  • 6 For $K=\min(x,t)-xt$ on $[0,1]$, compute the Rayleigh quotients of $u=1$ and $u=x(1-x)$ and compare with $\lambda_1=\frac1{\pi^2}$.
  • 7 Let $T$ be compact self-adjoint with nonzero eigenvalues $(\lambda_n)$ (with multiplicity). Prove that $T$ is Hilbert–Schmidt iff $\sum\lambda_n^2\lt\infty$, and compute $\sum\lambda_n^2$ for $K=\min(x,t)$.
  • 8 Prove that a compact self-adjoint $T$ with $T^k=0$ for some $k\ge1$ is $0$, and that $\|T^2\|=\|T\|^2$.
1 $Tu=\langle u,1\rangle+3x\langle u,x\rangle$. $T1=2$ and $Tx=3x\cdot\frac23=2x$; on $\{1,x\}^\perp$, $T=0$. So $\lambda=2$ with multiplicity $2$, and $0$. The projection onto $\operatorname{span}\{1,x\}$ is $Pu=\frac{\langle u,1\rangle}2+\frac{3x\langle u,x\rangle}2$, so $T=2P$ (checked numerically: eigenvalues $2,2,0,\dots$).
2 $\cos(x-t)=\cos x\cos t+\sin x\sin t$, and on $[0,2\pi]$ $\cos x\perp\sin x$ with $\|\cos\|^2=\|\sin\|^2=\pi$. So $T\cos=\pi\cos$, $T\sin=\pi\sin$: eigenvalue $\pi$ of multiplicity $2$, $\|T\|=\pi$.
3 $x\perp x^2$ on $[-1,1]$, $\|x\|^2=\frac23$, $\|x^2\|^2=\frac25$: $Tx=\frac23x$, $Tx^2=\frac25x^2$. Nonzero eigenvalues $\frac23,\frac25$; $\|T\|=\frac23$.
4 $\|T-T_N\|=\lambda_{N+1}=\frac4{(2N+1)^2\pi^2}$. This is $\approx0.0450,\ 0.0162,\ 0.00827$ for $N=1,2,3$. So $N=3$.
5 $\min(x,t)=\int_0^1\mathbf 1_{\{s\lt x\}}\mathbf 1_{\{s\lt t\}}ds$. By Fubini, for real $u$, $$\begin{aligned}\langle Tu,u\rangle&=\int_0^1\Big(\int_0^1\mathbf 1_{\{s\lt x\}}u(x)\,dx\Big)^2ds\\&=\int_0^1\Big(\int_s^1u\Big)^2ds\end{aligned}$$ (for complex $u$ use $|\cdot|^2$).
6 $u=1$: $\int\!\!\int(\min-xt)=\frac13-\frac14=\frac1{12}\approx0.0833$. $u=x(1-x)$: $R=\frac{17}{168}\approx0.10119$. Both are below $\frac1{\pi^2}\approx0.10132$, the second within $0.13\%$ (computed symbolically).
7 In an orthonormal basis of eigenvectors (Theorem 8.35(3)), $\sum\|Tf_m\|^2=\sum\lambda_n^2$, and the Hilbert–Schmidt norm does not depend on the basis (Definition 8.15). For $\min(x,t)$: $\sum\lambda_n^2=\int\!\!\int\min(x,t)^2=\frac16$.
8 $T^ke_n=\lambda_n^ke_n=0$ forces every $\lambda_n=0$, so $T=0$ by Theorem 8.35(1). $\|T^2\|=\|T^*T\|=\|T\|^2$ (§8.3); alternatively $T^2$ has eigenvalues $\lambda_n^2$, the largest $\lambda_1^2$.
IV
Integral equations and Sturm–Liouville problems.
  • 1 Solve $u(x)=1+\lambda\int_0^1xt\,u(t)\,dt$ for every real $\lambda$, and identify the characteristic value.
  • 2 Find the characteristic values of $u(x)=f(x)+\lambda\int_{-1}^1(1+xt)u(t)\,dt$, and the condition on $f$ for solvability at the smaller one.
  • 3 Find the eigenvalues and eigenfunctions of $-u''=\lambda u$, $u(0)=0$, $u'(1)=0$, and show that its Green's function is $\min(x,t)$.
  • 4 Use Mercer's theorem for $\min(x,t)$ to prove $\sum_{n\ge1}\frac1{(2n-1)^2}=\frac{\pi^2}8$.
  • 5 Use $\int\!\!\int G^2=\sum\mu_n^2$ for $G=\min(x,t)-xt$ to prove $\sum\frac1{n^4}=\frac{\pi^4}{90}$.
  • 6 Show that $0$ is an eigenvalue of $-u''=\lambda u$, $u'(0)=u'(1)=0$, so no Green's function exists. Instead find the Green's function of $-u''+u$ with these conditions and use it to prove that $\{1\}\cup\{\sqrt2\cos n\pi x\}_{n\ge1}$ is an orthonormal basis of $L^2[0,1]$.
  • 7 (Modelling) A taut string of unit length and unit tension, fixed at both ends, carries a load $f(x)$; its small deflection solves $-u''=f$, $u(0)=u(1)=0$. For the uniform load $f\equiv1$, expand the deflection in the eigenfunctions $\sqrt2\sin n\pi x$ and deduce $\sum_{k\ge0}\frac{(-1)^k}{(2k+1)^3}=\frac{\pi^3}{32}$.
1 $u=1+\lambda cx$ with $c=\int_0^1tu=\frac12+\frac{\lambda c}3$. For $\lambda\ne3$: $c=\frac3{2(3-\lambda)}$ and $u=1+\frac{3\lambda x}{2(3-\lambda)}$. For $\lambda=3$: $c=\frac12+c$ is impossible — no solution. $T$ has the single nonzero eigenvalue $\int_0^1t^2=\frac13$, so the characteristic value is $3$, and $f=1$ is not orthogonal to the eigenfunction $x$.
2 $T$ has eigenvalues $2$ (eigenfunction $1$) and $\frac23$ (eigenfunction $x$): characteristic values $\frac12$ and $\frac32$. At $\lambda=\frac12$ a solution exists iff $f\perp1$, i.e. $\int_{-1}^1f=0$.
3 $\lambda=\omega^2$: $u=\sin\omega x$, $\cos\omega=0$, so $\lambda_n=(n-\frac12)^2\pi^2$, $u_n=\sqrt2\sin((n-\frac12)\pi x)$; $\lambda\le0$ gives $u=0$. Green's function: $u_a=x$, $u_b=1$ (since $u_b'(1)=0$), $C=u_a'u_b-u_au_b'=1$: $G=\min(x,t)$, matching Example 4.
4 Example 4(d): $\sum\frac4{(2n-1)^2\pi^2}=\int_0^1x\,dx=\frac12$.
5 $\int_0^1\!\!\int_0^1(\min(x,t)-xt)^2\,dt\,dx=\frac1{90}$ (computed: by symmetry $2\int_0^1\!\int_0^xt^2(1-x)^2\,dt\,dx$). With $\mu_n=\frac1{n^2\pi^2}$: $\sum\frac1{n^4\pi^4}=\frac1{90}$.
6 $u\equiv1$ solves $-u''=0$ with $u'(0)=u'(1)=0$. For $L=-u''+u$: $u_a=\cosh x$ ($u_a'(0)=0$), $u_b=\cosh(1-x)$ ($u_b'(1)=0$), $C=\sinh x\cosh(1-x)+\cosh x\sinh(1-x)$ $=\sinh1$. So $$G(x,t)=\frac{\cosh(\min(x,t))\cosh(1-\max(x,t))}{\sinh1}.$$ $0$ is not an eigenvalue of $L$ ($\int(u'^2+u^2)=0$ forces $u=0$). Eigenfunctions of $Lu=\lambda u$ are $1$ and $\cos n\pi x$ with $\lambda=1+n^2\pi^2$; by Theorem 8.37 they form an orthonormal basis after normalising. (Numerically $T_G$ has eigenvalues $1,\ 0.0920,\ 0.0247,\dots=\frac1{1+n^2\pi^2}$.)
7 $u=T_G1$ with $G$ from Example 6: $u=\frac{x(1-x)}2$. With $e_n=\sqrt2\sin n\pi x$: $\langle1,e_n\rangle=\frac{\sqrt2(1-(-1)^n)}{n\pi}$, so $u=\sum\mu_n\langle1,e_n\rangle e_n=\sum_{n\text{ odd}}\frac{4}{n^3\pi^3}\sin n\pi x$. At $x=\frac12$: $\frac18=\frac4{\pi^3}\sum_{k\ge0}\frac{(-1)^k}{(2k+1)^3}$, giving $\frac{\pi^3}{32}\approx0.96895$ (series checked numerically). The coefficients decay like $n^{-3}$: the deflection is much smoother than the load.
🎯 GATE corner GATE MA · MCQGATE MA · MSQGATE MA · NAT

E1. GATE MA · MCQ Which one of the following operators on $\ell^2$ is compact? (A) $(x_1,x_2,x_3,\dots)\mapsto(x_2,x_3,x_4,\dots)$ (B) $(x_1,x_2,x_3,\dots)\mapsto(x_1,\tfrac{x_2}2,\tfrac{x_3}3,\dots)$ (C) $(x_1,x_2,x_3,\dots)\mapsto(x_1,0,x_3,0,x_5,\dots)$ (D) $(x_1,x_2,x_3,\dots)\mapsto\big((1+\tfrac1n)x_n\big)_n$

Idea Test $\|Te_n\|\to0$ (Theorem 8.30(1)); for diagonal operators this is also sufficient.
Options (A) $\|Le_{n+1}\|=1$; (C) $\|Pe_{2k-1}\|=1$ (a projection of infinite rank); (D) $\|Te_n\|=1+\frac1n\to1$. (B) is diagonal with $\frac1n\to0$.
Answer (B).

E2. GATE MA · MCQ Let $T:L^2[-1,1]\to L^2[-1,1]$, $(Tu)(x)=\int_{-1}^{1}(1+xt)\,u(t)\,dt$. Then $\sigma(T)$ is (A) $\{2,\tfrac23\}$ (B) $\{0,2,\tfrac23\}$ (C) $\{0,2\}$ (D) $[0,2]$

Idea $Tu=\langle u,1\rangle+x\langle u,x\rangle$ with $1\perp x$ on $[-1,1]$.
Compute $T1=2$, $Tx=\frac23x$, and $T=0$ on $\{1,x\}^\perp$ (infinite-dimensional). A compact operator on an infinite-dimensional space always has $0$ in its spectrum, and here $0$ is even an eigenvalue. Every other $\mu$ is in the resolvent set (Theorem 8.36(2)).
Answer (B). Option (A) forgets $0$; (D) confuses a compact operator with a multiplication operator.

E3. GATE MA · MSQ CSIR Part C Let $T\ne0$ be a compact self-adjoint operator on an infinite-dimensional separable Hilbert space $H$. Which of the following are necessarily true? (One or more options may be correct.) (A) $0\in\sigma(T)$. (B) $\|T\|$ or $-\|T\|$ is an eigenvalue of $T$. (C) Every eigenspace of $T$ is finite-dimensional. (D) $H$ has an orthonormal basis consisting of eigenvectors of $T$.

(A) True Theorem 8.29(5): a compact operator on an infinite-dimensional space is not invertible.
(B) True Theorem 8.34.
(C) False The eigenspace for $0$ can be infinite-dimensional: the rank-one operator $x\mapsto\langle x,e_1\rangle e_1$ has $\ker=\{e_1\}^\perp$. (Eigenspaces for nonzero eigenvalues are finite-dimensional.)
(D) True Theorem 8.35(3): the $e_n$ together with an orthonormal basis of $\ker T$ (countable, as $H$ is separable).
Answer (A), (B) and (D).

E4. GATE MA · MSQ Let $(Vu)(x)=\int_0^xu(t)\,dt$ on $L^2[0,1]$. Which of the following are true? (One or more options may be correct.) (A) $V$ is compact. (B) $V$ is self-adjoint. (C) $V$ has no eigenvalues. (D) $\|V\|$ equals the spectral radius of $V$.

(A) True Hilbert–Schmidt kernel, $\|V\|_{HS}=\frac1{\sqrt2}$.
(B) False $(V^*u)(x)=\int_x^1u$ (§8.3), which differs from $V$ (e.g. $V1=x$, $V^*1=1-x$).
(C) True Example 2(d).
(D) False $r(V)=0$, but $\|V\|=\frac2\pi$: $V^*V$ has kernel $1-\max(x,t)$, whose largest eigenvalue — from $-w''=\mu^{-1}w$, $w'(0)=0$, $w(1)=0$ — is $\frac4{\pi^2}$, so $\|V\|^2=\|V^*V\|=\frac4{\pi^2}$ (confirmed numerically).
Answer (A) and (C).

E5. GATE MA · NAT Let $(Tu)(x)=\int_0^1(1+2xt)\,u(t)\,dt$ on $L^2[0,1]$. The value of $\|T\|$, correct to three decimal places, is ______.

Idea Symmetric degenerate kernel: $\|T\|=\max|\lambda|$, and the nonzero eigenvalues come from a $2\times2$ matrix.
Matrix $Tu=\langle u,1\rangle+2x\langle u,t\rangle$. On $u=a+bx$: $T(a+bx)=\big(a+\tfrac b2\big)+\big(a+\tfrac{2b}3\big)x$, matrix $\begin{pmatrix}1&\frac12\\1&\frac23\end{pmatrix}$, trace $\frac53$, determinant $\frac16$. $\lambda^2-\frac53\lambda+\frac16=0$ gives $\lambda=\frac{5\pm\sqrt{19}}6\approx1.5598,\ 0.1069$.
Answer $\|T\|=\frac{5+\sqrt{19}}6\approx1.560$ (accept $1.559$ to $1.561$). The Hilbert–Schmidt norm $\|K\|_{L^2}=\sqrt{1+1+\frac49}\approx1.563$ and $\max|K|=3$ are only upper bounds — close here because $\lambda_2$ is small, but not equal.

E6. GATE MA · NAT Let $u\in C[0,1]$ satisfy $u(x)=x^2+2\int_0^1xt\,u(t)\,dt$ for all $x\in[0,1]$. The value of $u(1)$, correct to two decimal places, is ______.

Idea Degenerate kernel: $u=x^2+2cx$ with $c=\int_0^1tu(t)\,dt$. The characteristic value is $3$, so $\lambda=2$ gives a unique solution.
Compute $c=\int_0^1(t^3+2ct^2)\,dt=\frac14+\frac{2c}3$, so $\frac c3=\frac14$, $c=\frac34$, and $u(x)=x^2+\frac32x$.
Answer $u(1)=2.50$ (accept $2.49$ to $2.51$).
🧑‍🏫 For the teacher

Suggested plan — 3 lectures × 60 min.

  1. Compact operators. Definition 8.14, Theorem 8.29 (prove (2) and Riesz's lemma with (5)); Definition 8.15 and Theorem 8.30 (prove (1), (4), (5); sketch (2)). Example 1 as a class vote, then GATE corner E1.
  2. Spectra. Definition 8.16, Theorem 8.31 (Neumann series in full), Example 2 — draw the four spectra side by side. Definition 8.17 and Theorem 8.32 (the polarization proof of (2) is short and worth doing). State Theorem 8.33 and prove (3). Close with Example 7 (true/false) and E4.
  3. The spectral theorem and applications. Theorem 8.34 and Theorem 8.35 in full; run Interactive 1 with $\min(x,t)$ and then with $\cos\pi(x-t)$ to show a double eigenvalue. Examples 3 and 4; Interactive 2 to "see" $\|T-T_N\|=|\lambda_{N+1}|$. Theorem 8.36(2) via Example 5; Theorem 8.37 with Example 6. Drill E2, E5, E6.

Discussion questions. (1) The proof of Theorem 8.34 maximises $|\langle Tx,x\rangle|$. Where exactly does compactness enter, and what happens for $M_t$? (2) Why does a self-adjoint operator have $r(T)=\|T\|$, and why is this false for the Volterra operator? (3) A differential operator is unbounded, yet we proved completeness of its eigenfunctions with a theorem about bounded operators. How?

Misconceptions to watch for. "Compact means bounded"; "the spectrum is the set of eigenvalues"; "$0$ is an eigenvalue of every compact operator"; "every compact operator has an eigenvalue"; "$\|T\|=\max|\lambda|$" for non-symmetric kernels; confusing the characteristic value $\lambda$ of $u=f+\lambda Tu$ with the eigenvalue $1/\lambda$ of $T$; forgetting that the kernel of a degenerate-kernel operator is infinite-dimensional.

Suggested homework. I(1, 2, 3, 5, 6), II(1, 2, 4, 5, 8), III(1, 3, 4, 6, 7), IV(1, 2, 3, 6, 7); GATE corner E3, E5, E6.

🔬 Research–Teaching NexusResearch-tutored · optional, about 60–90 minutes

Research question. A random curve — a Brownian path $B(t)$ on $[0,1]$, whose covariance is $\operatorname{Cov}(B(s),B(t))=\min(s,t)$ — can be expanded in the eigenfunctions of its covariance operator (the Karhunen–Loève expansion). How many terms are needed to capture 95% and 99% of its total variance, and how good do the truncated random curves look?

Why it matters. Karhunen–Loève expansions are principal component analysis for functions. They are used to compress and denoise climate records and growth curves, to simulate random loads in structural engineering, to price path-dependent financial options, and to reduce the dimension of uncertain inputs in engineering simulations. The spectral theorem guarantees the expansion; the eigenvalues say how much each term is worth.

What to do.

  1. Using Example 4, write $B(t)=\sum_n\sqrt{\lambda_n}\,Z_n\,e_n(t)$ with independent standard normal $Z_n$, $\lambda_n=\frac4{(2n-1)^2\pi^2}$, $e_n=\sqrt2\sin((n-\frac12)\pi t)$. Show that the expected value of $\|B\|_2^2$ is $\sum\lambda_n=\frac12$ (Mercer).
  2. Compute the fraction $F_N=\sum_{n\le N}\lambda_n/\frac12$ and find the smallest $N$ with $F_N\ge0.95$ and with $F_N\ge0.99$.
  3. Discretise the kernel on a grid of 200 points (Nyström), compute the eigenvalues numerically and compare with the exact ones.
  4. Simulate a few Brownian paths with $N=5$ and $N=50$ terms (seeded random numbers) and compare them with a random-walk simulation. Repeat steps 2–3 for the Brownian bridge kernel $\min(s,t)-st$.

Tools: Python with NumPy (or MATLAB/Octave); a plotting tool.

Hand in: A table of $F_N$, the eigenvalue comparison, two plots of simulated paths, and a paragraph explaining to a climate scientist what the first eigenfunction "means".

Going further. Read about functional principal component analysis and about Mercer kernels in machine learning (kernel PCA, Gaussian processes), where the same spectral theorem is used with kernels such as $e^{-|s-t|^2}$.

Instructor note: $F_N=\frac8{\pi^2}\sum_{n\le N}(2n-1)^{-2}$, so $F_1\approx0.8106$, $F_2\approx0.9006$, $F_3\approx0.9331$, $F_4\approx0.9496$, $F_5\approx0.9596$: 95% needs $N=5$; $F_{20}\approx0.9899$, $F_{21}\approx0.9904$: 99% needs $N=21$, since $1-F_N\approx\frac2{\pi^2N}$. For the bridge, $\lambda_n=\frac1{n^2\pi^2}$, total $\frac16$, and $F_N=\frac6{\pi^2}\sum_{n\le N}n^{-2}$ ($F_1\approx0.608$). A 200-point midpoint Nyström discretisation reproduces $\lambda_1$ to about five significant figures.

🧪 Self-check quiz

0/5 correct

Q1. Which operator on $\ell^2$ is compact?

Q2. The spectrum of the diagonal operator $x\mapsto(x_n/n)$ on $\ell^2$ is

Q3. For a compact self-adjoint operator $T$, $\|T\|$ equals

Q4. The Volterra operator $Vu(x)=\int_0^xu$ on $L^2[0,1]$ is

Q5. The eigenfunctions $\sqrt2\sin n\pi x$ of $-u''=\lambda u$, $u(0)=u(1)=0$, form an orthonormal basis of $L^2[0,1]$ because