Source-linked AI summary
Mathematical and physical aspects of complex symmetric operators
Stephan Ramon Garcia, Emil Prodan, Mihai Putinar
TL;DR
The survey addresses the need to connect mathematical and physical perspectives on complex symmetric operators amid renewed interest in non-hermitian quantum mechanics. It synthesizes structural, extension-theoretic, spectral, and quantum-mechanical results, including criteria for real spectra and constructions of selfadjoint realizations. Its scope is bounded by the stated conditions for variational principles and by the lack of a settled functional model for irreducible complex symmetric operators.
Problem
The survey addresses the separation between mathematical treatments and the growing mathematical-physics literature on non-selfadjoint operators, including questions about representing observables and obtaining real spectra.
Method
It presents a non-technical survey connecting operator-theoretic results on complex symmetric operators with non-hermitian quantum mechanics and related examples.
Results
The survey develops results on operator structure, C-selfadjoint extensions, resolvents, spectral reality, C-orthonormal bases, and conjugate-linear symmetry, including sufficient criteria for real PT-symmetric spectra.
Takeaways & Limitations
Bounded C symmetries can realize PT-symmetric Hamiltonians as Hermitian operators under a new scalar product, whereas unbounded symmetries can permit multiple selfadjoint extensions.
Takeaways & Limitations
The variational principles require discreteness of the spectrum of |T| rather than compactness, and no concrete functional model for irreducible complex symmetric operators is yet clear.
Abstract
from arXiv · showhide
Recent advances in the theory of complex symmetric operators are presented and related to current studies in non-hermitian quantum mechanics. The main themes of the survey are: the structure of complex symmetric operators, $C$-selfadjoint extensions of $C$-symmetric unbounded operators, resolvent estimates, reality of spectrum, bases of $C$-orthonormal vectors, and conjugate-linear symmetric operators. The main results are complemented by a variety of natural examples arising in field theory, quantum physics, and complex variables.
1. INTRODUCTION
The survey responds to revived interest in non-hermitian quantum mechanics by connecting mathematical and physical work on complex symmetric operators. It situates the subject within operator theory, complex analysis, and diverse applications.
- Mathematical setting: Complex symmetric operators developed at the intersection of operator theory and complex analysis and became particularly relevant to truncated Toeplitz operators.
- Motivation: Recent non-hermitian quantum-mechanics research has renewed interest in the spectral analysis of complex symmetric operators.The survey aims to connect mathematicians and physicists while presenting the material at a non-technical level.
- Historical roots: The theory draws on classical work concerning complex symmetric matrices, automorphic functions, projective geometry, quadratic forms, symplectic geometry, and function theory.
- Extension theory: Complex symmetric differential operators have an extension theory connected naturally to symmetric operators in indefinite metric spaces.Both symmetries are expressed through relations of the form T ⊆ ST* S, with different involutions.
- Applications: Applications include quantum reaction dynamics, electric power modeling, high-voltage insulator simulation, transport, wave propagation, graph theory, scattering, and decay phenomena.
- Community bridge: The survey addresses a gap between the rapidly growing mathematical-physics literature and mathematical treatments of non-selfadjoint operators.A MathSciNet count reported more than 200 articles devoted solely to PT-symmetric operators.
Disclaimer.
The survey incorporates substantial material from the authors’ earlier publications while standardizing notation and presentation. It also records support, acknowledgments, and conventions used throughout the article.
- Disclaimer: The authors acknowledge extensive reuse of material from their previously published articles.
- Presentation: They state that the presentation and notation were streamlined and standardized throughout the survey.
- Notation: The article adopts customary mathematical notation, including inner products linear in the first slot.
- Notation: Vectors, matrices, operators, scalars, adjoints, transposes, unitary equivalence, and operator norms receive explicit notation conventions.
- Acknowledgments: The authors acknowledge constructive criticism, bibliographical guidance, and financial support from NSF and Nanyang Technological University.
2. COMPLEX SYMMETRIC OPERATORS
This section develops conjugations and complex symmetric operators as intertwined structures on Hilbert spaces. It connects operator symmetry with matrix representations and symmetric bilinear forms, while introducing examples and extension-theoretic context.
- 2.1. Conjugations: Conjugations generalize complex conjugation as conjugate-linear operators and provide the auxiliary symmetry underlying complex symmetric operators.
- 2.1. Conjugations: Conjugations appear in extension theory for unbounded symmetric operators and in von Neumann algebra theory.
- 2.1. Conjugations: Standard examples include pointwise conjugation on L2 spaces, Toeplitz and Szegő conjugations, symmetry-respecting conjugations, and PT conjugation.
- 2.1. Conjugations: For spin systems, time reversal is a conjugation for integer spin but an anti-conjugation for half-integer spin.
- 2.1. Conjugations: Every conjugation is unitarily equivalent to canonical conjugation and admits a conjugation-fixed orthonormal basis.Such vectors are called C-real, and the basis is a C-real orthonormal basis.
- 2.2. Complex symmetric operators: A bounded operator is complex symmetric exactly when it has a self-transpose matrix representation in some orthonormal basis.Equivalently, it is unitarily equivalent to a complex symmetric matrix on an appropriate ℓ2-space.
- 2.2. Complex symmetric operators: Every operator on a two-dimensional Hilbert space and every binormal operator is complex symmetric.
- 2.2. Complex symmetric operators: Every finite square matrix is similar to a complex symmetric matrix, which may retain any possible Jordan canonical form.
3. POLAR STRUCTURE AND SINGULAR VALUES
The section develops structural decompositions and variational descriptions for complex symmetric operators through conjugations, polar decomposition, and singular-value analysis. It also characterizes spectral information using conjugation-fixed approximate and exact antilinear eigenvectors.
- 3.1. Unitary decomposition: Every unitary operator factors as U = CJ for two conjugations, and the factorization makes U both C-symmetric and J-symmetric.The decomposition is not unique, as shown by finite-dimensional eigenvector constructions with offsetting unimodular parameters.
- 3.2. Refined polar decomposition: The refined polar decomposition writes a bounded C-symmetric operator as T = CJ|T|, with J commuting with |T| and its spectral projections.The construction extends a partial conjugation arising from the polar decomposition to a conjugation on the whole Hilbert space.
- 3.2. Refined polar decomposition: The refined decomposition supports applications to optimal approximation of Hankel operators and related extremal problems in function theory.
- 3.3. Approximate antilinear eigenvalue problems: For λ ≥ 0, λ belongs to σ(|T|) exactly when unit approximate eigenvectors can be chosen fixed by J.
- 3.3. Approximate antilinear eigenvalue problems: A nonnegative λ is an eigenvalue of |T| precisely when the associated antilinear eigenvalue problem has a nonzero J-fixed solution.
- 3.4. Variational principles: Complex symmetric matrices and compact C-symmetric operators admit minimax and variational principles for their singular values.These principles are linked to classical finite-dimensional factorizations and can support numerical computation of bound-state energies.
- 3.4. Variational principles: The variational proofs require discreteness of σ(|T|), not compactness itself, and apply to eigenvalues strictly above the essential spectrum.
4. SPECTRAL THEORY
The survey develops spectral structure for complex symmetric operators through direct-sum decompositions, Riesz idempotents, C-orthogonality, eigenstructure, and basis criteria.
- Direct sum decomposition: For dimensions at most 5, every complex symmetric operator is unitarily equivalent to a direct sum of irreducible complex symmetric operators.
- Direct sum decomposition: Finite-dimensional complex symmetric operators decompose into irreducible complex symmetric summands and blocks A⊕CA∗C with A irreducible and not complex symmetric.The block construction shows that complex symmetry need not pass to individual direct summands.
- Direct sum decomposition: In arbitrary dimensions, bounded complex symmetric operators decompose into completely reducible operators, irreducible operators, and blocks A⊕CA∗C.
- Riesz idempotents and orthogonality: Riesz idempotents associated with clopen spectral components are C-orthogonal, and generalized eigenspaces for distinct eigenvalues are mutually C-orthogonal.For compact C-symmetric operators, generalized eigenspaces are C-orthogonal as well.
- Eigenstructure: Eigenvectors corresponding to distinct eigenvalues are bilinearly orthogonal, while an isolated eigenvalue is simple exactly when it has no isotropic eigenvectors.The isolation hypothesis is essential: S⊕S∗ has simple eigenvalues throughout the open unit disk, yet every eigenvector is isotropic.
- C-orthonormal systems and Riesz bases: A complete C-orthonormal system is a Riesz basis precisely under equivalent boundedness conditions, with the optimal bound equal to the norm of A0.The equivalent conditions include being a Bessel sequence, having Riesz bounds M−1 and M, and boundedness of the associated matrix operator.
- C-orthonormal systems and Riesz bases: For pure dissipative C-symmetric operators with simple spectrum and separated complete eigenvectors, the eigenvectors form a Riesz basis and support norm-convergent skew Fourier expansions.For the bounded setting, the construction also yields a positive C-orthogonal extension and a C-real orthonormal basis.
- Further spectral structure: Bounded C-symmetric operators are decomposable exactly when they satisfy Bishop’s condition (β), and C-symmetric operators can be connected continuously to diagonal models in the stated construction.The deformation uses C-orthogonal exponentials and remains norm continuous through intermediate C-symmetric operators.
5. UNBOUNDED COMPLEX SYMMETRIC OPERATORS
The section develops unbounded complex symmetric operator theory, emphasizing C-selfadjoint extensions, resolvent criteria, compact-resolvent eigenfunction bases, and polar decomposition. It also illustrates these results with differential and Schrödinger operators.
- Basic definitions: C-symmetry for unbounded operators is defined by T ⊆ CT*C, while C-selfadjointness strengthens this to T = CT*C.The distinction requires care because “symmetric” has different meanings for matrices and unbounded operators.
- Examples and bases: Examples include Sturm–Liouville, Schrödinger, and dissipative operators, while eigenfunctions may form a Riesz basis in some cases but fail to form any Hilbert, Riesz, or Schauder basis in others.Scaled Schrödinger Hamiltonians are C-selfadjoint, and certain operators have complete eigenfunctions by Keldysh’s theorem.
- C-selfadjoint extensions: Every densely defined C-symmetric operator admits a C-selfadjoint extension, without requiring C-reality or dissipativity.This generalizes earlier extension results that imposed additional hypotheses.
- Spectral structure: For unbounded C-selfadjoint operators with compact resolvent, an orthonormal basis consists of solutions to the associated antilinear eigenvalue problem.The result supports resolvent-norm estimates and is complemented by a corollary involving the positive antilinear eigenvalues.
- Refined polar decomposition: When zero lies in the resolvent, a densely defined C-selfadjoint operator has the refined polar decomposition T = CJ|T|, with |T| positive selfadjoint and J a commuting conjugation.The converse also holds for operators having the stated decomposition properties.
- C-selfadjoint extensions: A C-symmetric operator is C-selfadjoint when ran(T − λ)D(T) = H for some complex λ, with related criteria expressed through resolvent conditions.The section also gives a domain decomposition criterion and notes that regular resolvent points need not exist in general.
6. PT -SYMMETRIC HAMILTONIANS
The section surveys PT-symmetric Hamiltonians, emphasizing criteria for real spectra and the additional requirements for a probabilistically interpretable quantum theory. It discusses rigorous examples, perturbative approaches, similarity to selfadjoint operators, and the role of boundary conditions.
- Spectral reality: PT symmetry is necessary for real energy eigenvalues in the discussed finite-dimensional setting, but it is generally not sufficient.
- Spectral reality: For diagonalizable finite-dimensional PT-symmetric Hamiltonians, commuting involutions determine whether all eigenvalues are real or whether a conjugate complex pair occurs.
- Spectral reality: For a nondiagonalizable Jordan block with one eigenvector, PT symmetry yields real eigenvalues, whereas failure of PT symmetry yields complex eigenvalues.
- Physical interpretation: Reality of the spectrum is insufficient for probabilistic quantum mechanics because unitary dynamics also requires a suitable positive-definite inner product.
- Physical interpretation: A bounded symmetry C permits a new scalar product making the Hamiltonian Hermitian, while an unbounded C generally leads to multiple selfadjoint extensions in a larger space.
- Examples: The perturbed cubic oscillator has a spectrum whose reality was conjectured, numerically supported, and later established rigorously using complex-domain PDE techniques.
- Similarity and bases: The survey also reports that the perturbed cubic oscillator is not similar to a selfadjoint operator and its eigenfunctions do not form a Riesz basis.
- Differential operators: For the PT-symmetric differential-operator class discussed, the spectrum is entirely real when β ≥ 0, whereas complex eigenvalues may appear when β < 0.
7. MISCELLANEOUS APPLICATIONS
The survey develops applications of complex symmetric and related non-selfadjoint operators across quantum models, operator theory, harmonic analysis, and complex variables. These applications yield criteria for spectral reality, decay estimates, functional models, extremal bounds, and structural results for boundary operators.
- Quantum and spectral applications: Magnetic Schrödinger Hamiltonians in weak-field insulating regimes have a spectral gap, motivating sharp exponential decay estimates for their resolvents.The non-selfadjoint scaled Hamiltonians are obtained by conjugating the magnetic Hamiltonian with a bounded invertible transformation that leaves its domain unchanged.
- Conjugate-linear operator models: Conjugate-linear R-selfadjoint operators with a cyclic vector receive a functional model, and a spectral mapping theorem holds for specified continuous functions of bounded R-selfadjoint operators.The spectral mapping result applies when the polar decomposition has a conjugation commuting with the modulus.
- Bilinear forms and complex variables: The second singular value gives the best possible codimension-one bound for the bilinear form and the optimal L2 bound on harmonic conjugation.The associated variational statements identify σ2 as the optimal constant and establish its sharpness on a hyperplane through the origin.
- Boundary operators: In planar boundary and Bergman-space problems, compact conjugate-linear eigenvalue problems describe the Neumann-Poincaré spectrum, whose nontrivial spectrum is origin-symmetric and shared by interior and exterior operators.The eigenfunctions are orthogonal and complete, and the conjugate-linear transform is an isometric isomorphism on the relevant Bergman spaces.
- Symmetrizable operators: Symmetrizable compact operators have real spectrum, no nontrivial Jordan chains at nonzero eigenvalues, and adjoint eigenvectors spanning the space; this applies to the Neumann-Poincaré and Beurling transforms.The stated hypothesis is the existence of a strictly positive bounded operator A satisfying RM = M*R.