Spectral Theory of Compact Operators

Compact operators on a Banach (or Hilbert) space have spectra that behave almost like those of matrices: the nonzero spectrum consists of at most countably many eigenvalues of finite multiplicity, accumulating only at

0

. The central result is the Riesz–Schauder theory and, for self-adjoint compact operators, the spectral theorem giving an orthonormal eigenbasis. Applies throughout integral equations, PDE eigenvalue problems, and quantum mechanics. Developed by Fredholm, Hilbert, Schmidt, F. Riesz, and Schauder (1900–1930).

A bounded linear operator $T$ on a Banach space $X$ is compact if it maps bounded sets to relatively compact sets — equivalently, every bounded sequence $(x_{n})$ has a subsequence with $(T x_{n})$ convergent. Such operators are the infinite-dimensional analogue of finite-rank matrices, and their spectral structure is correspondingly tame: outside any disk around $0$ the spectrum is finite. This matters because countless problems in analysis — solving $T u - λ u = f$ , diagonalizing integral kernels, finding vibration modes — reduce to compact operators where a clean eigenvalue calculus survives. The surprising fact: even when $T$ is injective with dense range, $0$ may still be a spectral point that is not an eigenvalue, so the only "essential" spectral behaviour collapses entirely onto the single point $0$ .

The Core Insight

Compactness forces the operator to "almost factor through finite dimensions." The Fredholm alternative for $T - λ I$ ( $λ \neq = 0$ ) then holds verbatim as in linear algebra: either $T - λ I$ is invertible, or $λ$ is an eigenvalue with finite-dimensional eigenspace and finite Riesz index. The infinitude of $X$ is squeezed into a single accumulation point at the origin.

Formal Statement

Compact Operator

Let $X, Y$ be Banach spaces. A linear map $T : X \to Y$ is compact if $\overline{T (B_{X})}$ is compact in $Y$ , where $B_{X} = {x : ∥ x ∥ \leq 1}$ . Equivalently, $T$ is compact iff for every bounded sequence $(x_{n}) \subset X$ , the sequence $(T x_{n})$ has a convergent subsequence. We write $T \in K (X, Y)$ .

Spectrum and Its Parts

For a bounded operator $T$ on a complex Banach space $X$ , the resolvent set is

ρ (T) = {λ \in C : (T - λ I)^{- 1} \in B (X)},

⟳

λ \in ρ (T)

iff

T - λ I

is a bijection of

X

; boundedness of the inverse is then automatic by the open mapping theorem.

and the spectrum is $σ (T) = C ∖ ρ (T)$ . It decomposes as

σ (T) = σ_{p} (T) \cup σ_{c} (T) \cup σ_{r} (T),

⟳classify

T - λ I

by injectivity and density/closedness of range — point spectrum (not injective), continuous spectrum (injective, dense non-closed range), residual spectrum (injective, non-dense range).

where $σ_{p}$ (point/eigenvalues), $σ_{c}$ (continuous), $σ_{r}$ (residual).

Riesz–Schauder Spectral Theorem for Compact Operators

Let $X$ be an infinite-dimensional complex Banach space and $T \in K (X)$ . Then:

$0 \in σ (T)$ .
$σ (T) ∖ {0}$ consists entirely of eigenvalues.
Each nonzero $λ \in σ (T)$ has finite algebraic and geometric multiplicity:

dim ker (T - λ I)^{k} < \infty for all k \geq 1,

⟳

ker (T - λ I)^{k}

is a closed subspace on which

T

acts as

λ I

+ nilpotent; compactness restricted to it forces finite dimension since the identity on it must be compact.

$σ (T)$ is at most countable, and its only possible accumulation point is $0$ :

\forall ε > 0, # {λ \in σ (T) : ∣ λ ∣ \geq ε} < \infty.

⟳an infinite sequence of distinct eigenvalues bounded away from

0

produces, via near-orthonormal eigenvectors and Riesz's lemma, a bounded sequence whose images have no convergent subsequence — contradicting compactness.

Fredholm Alternative

For $T \in K (X)$ and $λ \neq = 0$ , exactly one holds:

(a) $T - λ I$ is bijective (so $λ \in ρ (T)$ and the equation $(T - λ I) u = f$ has a unique solution for every $f$ );

(b) $ker (T - λ I) \neq = {0}$ , with

dim ker (T - λ I) = dim ker (T^{*} - \overset{ˉ}{λ} I) = codim ran (T - λ I) < \infty.

⟳

T - λ I = - λ (I - λ^{- 1} T)

is a "compact perturbation of the identity"; its index is

0

by the Riesz theory, equating kernel and cokernel dimensions, with solvability of

(T - λ I) u = f

governed by the orthogonality

⟨ f, w ⟩ = 0

for all

w \in ker (T^{*} - \overset{ˉ}{λ} I)

Hilbert–Schmidt / Spectral Theorem for Compact Self-Adjoint Operators

Let $H$ be a separable Hilbert space and $T = T^{*} \in K (H)$ . Then there exist real eigenvalues $(λ_{n})$ with $∣ λ_{1} ∣ \geq ∣ λ_{2} ∣ \geq \dots \to 0$ and an orthonormal system $(e_{n})$ of eigenvectors, $T e_{n} = λ_{n} e_{n}$ , such that

T x = n \sum λ_{n} ⟨ x, e_{n} ⟩ e_{n} \forall x \in H,

⟳existence of a largest eigenvalue follows from

∥ T ∥ = sup_{∥ x ∥ = 1} ∣ ⟨ T x, x ⟩ ∣

being attained by a compactness/weak-limit argument; iterate on the orthogonal complement and use that eigenvectors for distinct eigenvalues are orthogonal.

and the $(e_{n})$ together with an orthonormal basis of $ker T$ form an orthonormal basis of $H$ . Moreover $∥ T ∥ = max_{n} ∣ λ_{n} ∣ = ∣ λ_{1} ∣$ .

Why It Holds

Riesz Theory: discreteness of nonzero spectrum

Let $T \in K (X)$ , $λ \neq = 0$ . Set $S = T - λ I$ .

Step 1 — Kernel is finite-dimensional. On

N := ker S = ker (T - λ I)

we have

T x = λ x

, so

T ∣_{N} = λ I_{N}

. Since

T

is compact and

λ \neq = 0

, the identity

I_{N} = λ^{- 1} T ∣_{N}

is compact, forcing

dim N < \infty

(a normed space with compact unit ball is finite-dimensional, by Riesz's lemma). Step 2 — Range is closed. Write

X = N \oplus M

with

M

closed complement (possible since

N

is finite-dimensional). The restriction

S ∣_{M} : M \to ran S

is injective; one shows

∥ S x ∥ \geq c ∥ x ∥

M

for some

c > 0

, otherwise a unit sequence

x_{n} \in M

with

S x_{n} \to 0

gives, via compactness of

T

, a convergent subsequence

x_{n_{k}} \to x \in M

with

S x = 0

, contradicting injectivity on

M

. Hence

ran S

is closed. Step 3 — Finiteness of the spectrum off any disk. Suppose

(λ_{n})

are distinct eigenvalues with

∣ λ_{n} ∣ \geq ε > 0

and eigenvectors

u_{n}

. The

u_{n}

are linearly independent; let

X_{n} = span (u_{1}, \dots, u_{n})

, a strictly increasing chain of closed subspaces. By Riesz's lemma pick

y_{n} \in X_{n}

∥ y_{n} ∥ = 1

dist (y_{n}, X_{n - 1}) \geq \frac{1}{2}

. For

m < n

one computes

∥ T y_{n} - T y_{m} ∥ = ∥ λ_{n} y_{n} - (λ_{n} y_{n} - T y_{n} + T y_{m}) ∥ \geq \frac{∣ λ _{n} ∣}{2} \geq \frac{ε}{2},

⟳

(λ_{n} I - T) y_{n}

and

T y_{m}

both lie in

X_{n - 1}

, while

λ_{n} y_{n}

projects to

\geq 1/2

distance from

X_{n - 1}

; combine with

T

acting triangularly on the chain.

so $(T y_{n})$ has no Cauchy subsequence, contradicting compactness. Thus only finitely many $∣ λ_{n} ∣ \geq ε$ .

Step 4 — Index zero / Fredholm. The ascending kernels

ker S^{k}

stabilize at finite index

p

(the Riesz index), and

X = ker S^{p} \oplus ran S^{p}

. On the finite-dimensional

ker S^{p}

S

is nilpotent; on

ran S^{p}

S

is bijective. This yields

dim ker S = codim ran S

, the Fredholm alternative.

■

Spectral theorem for compact self-adjoint $T$

Step 1 — Largest eigenvalue. For self-adjoint

T

∥ T ∥ = sup_{∥ x ∥ = 1} ∣ ⟨ T x, x ⟩ ∣

. Take

∥ x_{n} ∥ = 1

with

⟨ T x_{n}, x_{n} ⟩ \to λ_{1}

where

∣ λ_{1} ∣ = ∥ T ∥

. By compactness

T x_{n} \to y

(subsequence) and one shows

∥ T x_{n} - λ_{1} x_{n} ∥^{2} = ∥ T x_{n} ∥^{2} - 2 λ_{1} ⟨ T x_{n}, x_{n} ⟩ + λ_{1}^{2} \to 0

, so

x_{n} \to λ_{1}^{- 1} y =: e_{1}

is a unit eigenvector with

T e_{1} = λ_{1} e_{1}

(assuming

T \neq = 0

). Step 2 — Iterate. The orthogonal complement

{e_{1}}^{⊥}

T

-invariant (self-adjointness), and

T ∣_{{e_{1}}^{⊥}}

is again compact self-adjoint with norm

∣ λ_{2} ∣ \leq ∣ λ_{1} ∣

. Repeat to extract

e_{2}, e_{3}, \dots

. Step 3 — Convergence to zero. If infinitely many

λ_{n}

occur,

∣ λ_{n} ∣ = ∥ T ∣_{{e_{1}, \dots, e_{n - 1}}^{⊥}} ∥ \to 0

, since otherwise Step 3 of the Riesz proof is violated. Step 4 — Completeness on $(ker T)^{⊥}$ . For

x \in H

x - \sum_{n} ⟨ x, e_{n} ⟩ e_{n}

lies in

⋂_{n} {e_{n}}^{⊥}

where

T

has norm

\to 0

, hence

T

vanishes there: the residual lies in

ker T

. Equation [6] follows.

■

David Hilbert | 1862–1943

Hilbert's six papers on integral equations (1904–1910) introduced the spectral viewpoint that now bears his name. He treated kernels $K (s, t)$ as infinite "matrices," developed eigenvalue expansions, and coined the language of bilinear and quadratic forms in infinitely many variables. His student Erhard Schmidt then geometrized the theory into the language of orthonormal systems and what we now call Hilbert spaces, making the compact self-adjoint spectral theorem a clean geometric statement.

Building Intuition

Think of a compact operator as a matrix whose entries decay fast enough that it is "almost finite." A diagonal operator $T e_{n} = λ_{n} e_{n}$ on $ℓ^{2}$ is compact iff $λ_{n} \to 0$ — the eigenvalues must drain toward zero. The accumulation at $0$ is unavoidable: there simply is not "room" in a bounded set's image to keep large eigenvalues spread out, because compactness demands convergent subsequences. The single point $0$ is where all the infinite-dimensional complexity gets dumped; everything else looks like ordinary linear algebra, complete with eigenvalues, multiplicities, and a Jordan-like structure (the Riesz index $p$ ).

Spectrum of a compact operator: eigenvalues accumulating only at the origin

Worked Examples

Diagonal compact operator on $\ell^2$

Given.

T : ℓ^{2} \to ℓ^{2}

T (x_{1}, x_{2}, x_{3}, \dots) = (x_{1}, \frac{x _{2}}{2}, \frac{x _{3}}{3}, \dots)

, i.e.

T e_{n} = \frac{1}{n} e_{n}

. Step 1 — Compactness. The finite-rank truncations

T_{N} (x) = (\frac{x _{1}}{1}, \dots, \frac{x _{N}}{N}, 0, \dots)

satisfy

∥ T - T_{N} ∥ = \frac{1}{N + 1} \to 0

. A norm-limit of finite-rank operators is compact, so

T \in K (ℓ^{2})

. Step 2 — Eigenvalues.

σ_{p} (T) = {1, \frac{1}{2}, \frac{1}{3}, \dots}

, each simple, with eigenvector

e_{n}

. Step 3 — Spectrum. The accumulation point

0

must be included:

σ (T) = {\frac{1}{n} : n \geq 1} \cup {0}

. Note

0

is not an eigenvalue (

T

is injective) but lies in

σ_{c} (T)

since

ran T

is dense but not closed. Answer.

σ (T) = {1, \frac{1}{2}, \frac{1}{3}, \dots} \cup {0}, 0 \in σ_{c} (T), ∥ T ∥ = 1.

Solving an integral equation via the Fredholm alternative

Given. The Fredholm equation

u (s) - μ \int_{0}^{1} s t u (t) d t = f (s)

L^{2} [0, 1]

, with kernel

K (s, t) = s t

. Step 1 — Identify the operator.

T u (s) = \int_{0}^{1} s t u (t) d t = s ⟨ u, t ⟩_{L^{2}}

. This is rank one (range

= span {s}

), hence compact and self-adjoint. Step 2 — Eigenvalue.

T u = λ u

forces

u (s) = c s

. Then

T (s) = s \int_{0}^{1} t \cdot t d t = s \cdot \frac{1}{3}

. So the single nonzero eigenvalue is

λ_{1} = \frac{1}{3}

with eigenvector

e_{1} (s) = 3 s

. Step 3 — Apply Fredholm alternative for $μ T u - u$ . We solve

(I - μ T) u = f

. This is uniquely solvable iff

\frac{1}{μ} \in / σ_{p} (T) ∖ {0}

, i.e. iff

μ \neq = \frac{1}{λ _{1}} = 3

. Step 4 — Explicit solution for $μ \neq = 3$ . Write

u = f + μ s c

with

c = ⟨ u, t ⟩

. Then

c = ⟨ f, t ⟩ + μ c ⟨ s, t ⟩ = ⟨ f, t ⟩ + \frac{μ}{3} c

, giving

c = \frac{⟨ f , t ⟩}{1 - μ /3}

. Answer.

u (s) = f (s) + \frac{μ s}{1 - μ /3} \int_{0}^{1} t f (t) d t, μ \neq = 3;

for

μ = 3

the equation is solvable iff

\int_{0}^{1} t f (t) d t = 0

(Fredholm orthogonality).

Common Misconceptions

" $0$ is always an eigenvalue of a compact operator." False.

0

always lies in the spectrum (in infinite dimensions), but it need not be an eigenvalue: the diagonal operator

T e_{n} = \frac{1}{n} e_{n}

is injective, so

ker T = {0}

, yet

0 \in σ_{c} (T)

. The statement "

0 \in σ (T)

" holds because a compact operator on an infinite-dimensional space cannot be invertible — its inverse would make the identity compact. "Compact operators always have at least one eigenvalue." False in general Banach (and even Hilbert) spaces if

T

is not self-adjoint or normal. The nonzero spectrum can be empty — see the counterexample below. The spectral theorem's guaranteed eigenbasis requires self-adjointness (or normality, via

σ_{p}

C

Compact operator with no nonzero eigenvalues

The Volterra operator on $L^{2} [0, 1]$ ,

(V u) (s) = \int_{0}^{s} u (t) d t,

⟳

V

is Hilbert–Schmidt with kernel

K (s, t) = 1_{{t \leq s}}

, hence compact; a suspected eigenfunction

V u = λ u

with

λ \neq = 0

would be differentiable with

u^{'} = λ^{- 1} u

u (0) = 0

, forcing

u \equiv 0

is compact but quasinilpotent: $σ (V) = {0}$ and it has no eigenvalues at all. This shows the Riesz theory guarantees only that the nonzero spectrum consists of eigenvalues — it does not guarantee any exist. ( $V$ is non-self-adjoint; $V + V^{*}$ is the rank-one projection-like operator with full real eigenvalue structure.)

Connections

The compact spectral theorem sits inside a hierarchy. It is the cleanest case of the Spectral Theorem for Bounded Self-Adjoint Operators, which for non-compact operators replaces the discrete sum [6] by an integral against a Projection-Valued Measure. Compact operators form a closed two-sided ideal in $B (X)$ ; quotienting yields the Calkin Algebra, where invertibility encodes Fredholm Operators of arbitrary index — the Riesz theory is exactly the index-zero, compact-perturbation case. On Hilbert space the compact operators are the closure of finite-rank operators, and the refined classes (trace class, Hilbert–Schmidt Operators, Schatten $p$ -classes) are defined by summability of the singular values $s_{n} (T)$ — the eigenvalues of $∣ T ∣ = (T^{*} T)^{1/2}$ from the Singular Value Decomposition.

Real-World Applications

Sturm–Liouville theory & PDE eigenvalues. Inverting an elliptic operator $- Δ$ on a bounded domain gives a compact self-adjoint Green's operator; its discrete eigenvalue expansion [6] yields the vibration modes of drums and the energy levels of confined quantum systems.
Quantum mechanics — bound states. The resolvent of a Schrödinger operator with confining potential is compact; the discrete spectrum (eigenvalues $\to$ accumulation) gives the bound-state energies, while the continuous spectrum lives elsewhere.
Integral equations (Fredholm). Scattering theory, potential theory, and boundary-integral methods reduce to $(I - K) u = f$ with $K$ compact; the Fredholm alternative dictates solvability and is the backbone of the boundary element method.
Image compression & PCA. The SVD of compact integral operators / large matrices, with rapidly decaying singular values, justifies low-rank truncation — the quantitative form of "compact = almost finite-rank."
Tikhonov regularization of inverse problems. Ill-posed inverse problems $T x = y$ with compact $T$ have $T^{- 1}$ unbounded (because $s_{n} \to 0$ ); regularization filters the small singular values to stabilize reconstruction in tomography and deblurring.
Statistical learning — kernel methods. Mercer's theorem expands a positive-definite kernel as $K (s, t) = \sum_{n} λ_{n} e_{n} (s) e_{n} (t)$ via the compact self-adjoint integral operator; this underlies Gaussian processes and the spectral analysis of RKHS.
Markov chains & transfer operators. Compact transfer (Perron–Frobenius) operators in dynamical systems have a spectral gap whose subleading eigenvalues govern mixing and decay of correlations.

Historical Development

1900–1903 — Ivar Fredholm solves the integral equation $u - λ K u = f$ using infinite determinants, establishing the "Fredholm alternative" for continuous kernels.
1904–1910 — David Hilbert publishes his series on integral equations, introducing eigenvalue expansions and the theory of infinitely many variables (quadratic forms, "vollstetig" = completely continuous operators).
1907 — Erhard Schmidt recasts Hilbert's theory geometrically with orthonormal systems, proving the eigenfunction expansion (the Hilbert–Schmidt expansion).
1909 — Frigyes Riesz and Maurice Fréchet independently prove the $L^{2}$ representation theorem, anchoring Hilbert-space methods.
1916 — Frigyes Riesz develops the abstract theory of completely continuous (compact) operators on $C [a, b]$ , proving discreteness of the spectrum and finite multiplicity — the Riesz theory in essentially modern form.
1918 — Riesz proves what is now Riesz's lemma, the key compactness tool for the spectral discreteness argument.
1930 — Juliusz Schauder extends the Fredholm alternative to general Banach spaces and proves compactness of the adjoint (Schauder's theorem: $T$ compact $⟺ T^{*}$ compact).
1932 — John von Neumann's Mathematical Foundations of Quantum Mechanics and Stone's work place the spectral theorem on rigorous Hilbert-space footing, distinguishing compact (discrete) from continuous spectra.
1940s–1950s — Development of Schatten–von Neumann classes (trace class, Hilbert–Schmidt) quantifies compactness by singular-value summability.

Subtleties & Limits

Where the clean theory ends. Several genuinely hard or open issues lurk:

Invariant subspace problem. Whether every bounded operator on a separable infinite-dimensional Hilbert space has a nontrivial closed invariant subspace remains open. For compact operators the answer is famously yes — Aronszajn–Smith (1954), extended by Lomonosov (1973) to operators commuting with a nonzero compact operator — making compactness a rare island of resolution within an open ocean.
Eigenvalue vs. singular-value asymptotics. The relation between the eigenvalues $λ_{n} (T)$ and singular values $s_{n} (T)$ is subtle: Weyl's inequality $\sum ∣ λ_{n} ∣^{p} \leq \sum s_{n}^{p}$ holds, but precise asymptotics of eigenvalues for non-self-adjoint compact operators (Volterra-type, pseudodifferential) remain delicate and active.
Completeness of eigenvectors for non-self-adjoint compact operators is not guaranteed; root-vector completeness (Keldysh-type theorems) requires extra structure and is sharp only under decay conditions on $s_{n}$ .
Non-quasi-analytic perturbations. Small non-compact perturbations can dissolve the discrete spectrum into continuous spectrum, and pinpointing the essential spectrum under such perturbations connects to deep Fredholm-theoretic stability questions.

Spectral Theory of Compact Operators

The Core Insight

Formal Statement

Compact Operator

Spectrum and Its Parts

Riesz–Schauder Spectral Theorem for Compact Operators

Fredholm Alternative

Hilbert–Schmidt / Spectral Theorem for Compact Self-Adjoint Operators

Why It Holds

Riesz Theory: discreteness of nonzero spectrum

Spectral theorem for compact self-adjoint $T$

David Hilbert | 1862–1943

Building Intuition

Spectrum of a compact operator: eigenvalues accumulating only at the origin

Worked Examples

Diagonal compact operator on $\ell^2$

Solving an integral equation via the Fredholm alternative

Common Misconceptions

Compact operator with no nonzero eigenvalues

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also

Before this

Where this leads