Skip to content

States, Hadamard Structure, and Microlocal Control

A physically useful quantum state must be positive and have suitable short-distance behavior, but these conditions rarely select it uniquely. This chapter separates state construction, representation, ultraviolet admissibility, and physical selection.

Begin with Vacuum Ambiguity, Time Flow, and Observer Dependence to identify the time-flow assumptions behind positive frequency and a vacuum.

Helpful background. Singular Support and Wavefront Sets and Distributional Kernels and Distributions on Manifolds develop the definitions used by the chapter’s opening pages.

  1. Vacuum Ambiguity, Time Flow, and Observer Dependence — Explains why a generic background does not determine positive frequency and identifies the extra structure behind a vacuum claim.
  2. Complex Structures and One-Particle Spaces — Derives a one-particle Hilbert space from a positive symplectic-compatible complex structure.
  3. Quasifree States and Two-Point Functions — Gives the field-equation, commutator, positivity, reality, and continuity tests on two-point data.
  4. Bogoliubov Transformations and Unitary Implementability — Separates canonical mode mixing from its stronger Hilbert–Schmidt implementability condition.
  5. GNS Representations, Local Normality, and Local Quasiequivalence — Explains why globally inequivalent representations can describe the same local normal states.
  6. Hadamard Parametrix and Short-Distance Structure — Separates universal geometric singular terms from the smooth state-dependent remainder.
  7. Hadamard Admissibility and the Two-Point Wavefront Criterion — States the future-directed null wavefront condition and its limits.
  8. Propagation of the Hadamard Property — States the hyperbolic theorem that carries suitable local Cauchy data through the full development.
  9. Constructing Hadamard States by Deformation and Gluing — Turns that theorem into a state-construction method while retaining positivity and nonuniqueness checks.
  10. Adiabatic States, WKB Order, and Regularity — Relates high-momentum WKB order to regularity without confusing state choice with subtraction.
  11. Ground, KMS, and Symmetry-Selected States — Identifies when a normalized time flow or symmetry distinguishes a state and when obstructions prevent it.
  12. States of Low Energy and Smeared-Energy Selection — Explains observer- and sampling-dependent minimization in cosmological settings.
  13. Hadamard States for Fermion and Gauge Fields — Adapts microlocal and positivity tests to CAR systems and constrained gauge fields.
  14. Instantaneous Vacua, Alpha Vacua, Sorkin–Johnston States, and Failure Tests — Compares prominent prescriptions against ultraviolet, positivity, locality, and regulator-removal requirements.

For a quasifree field, two-point data determine higher correlations, subject to positivity and the canonical algebra. The Hadamard condition further restricts short-distance singularities so that local composite observables can be renormalized; its wavefront-set formulation exposes the causal directions of those singularities Radzikowski 1996, pp. 529–553.

Hadamard admissibility does not minimize energy or establish thermality. Ground, KMS, adiabatic, and low-energy constructions use different additional information. Comparing them requires keeping the representation and limiting procedure fixed rather than treating every acceptable state as the same vacuum.

The comparison below separates the assumptions and tests needed for these different conclusions.

State classExistence or construction dataUniquenessSymmetry or selectorUltraviolet regularityLocal comparisonDecisive failure or downgrade
Scalar quasifreePositive covariance, field equation, reality, continuity, and antisymmetric part −iE-iEGenerally many; purity is a separate conditionNone requiredUnrestricted until an adiabatic or Hadamard condition is addedMust be tested from restricted covariancesOne negative quadratic form removes statehood
Scalar HadamardPositive bisolution with the future-null two-point wavefront relationHighly nonunique; smooth bisolutions remainNone requiredFull Hadamard classQuasifree Hadamard representations are locally quasiequivalent under the standard hypothesesExtra wavefront directions or a nonsmooth difference from a Hadamard reference
Finite-order adiabaticNormalized WKB Cauchy data or the stated Sobolev wavefront boundNonunique; depends on initialization and order conventionUsually exploits spatial homogeneityControlled only through the declared finite orderAvailable only above the regularity thresholds of the relevant theoremTreating the order label as full Hadamard form or as an observable subtraction
GroundA complete stationary evolution with nonnegative GNS generatorSometimes unique within a stated regular quasifree classInvariant under the chosen normalized time flowHadamard under additional static or stationary spectral hypothesesAs for the resulting Hadamard representationZero modes, lack of a lower spectral bound, ergoregions, or superradiance
KMSA star-automorphism flow and KMS strip analyticity at declared β\betaThermal phases can be nonuniqueEquilibrium relative to that flowHadamard under the applicable stationary free-field hypothesesPhase and sector must be declared; Hadamard quasifree cases have the usual local comparisonRescaling the flow, infrared divergence, or incompatible horizon conditions
Euclidean or Bunch–DaviesAnalytic continuation and regular Euclidean data in the supported de Sitter field sectorDistinguished in the relevant invariant Hadamard class, not for every mass or couplingFull de Sitter invarianceHadamardLocally comparable to other Hadamard statesThe massless minimally coupled zero mode or replacement by a non-Euclidean alpha state
State of low energyA Robertson–Walker trial class, observer, and smooth compact sampling functionA minimizer relative to those declared inputs, not universallySampled-energy selection, not generally stationaryHadamard in Olbermann’s settingThe usual Hadamard local comparison appliesPointlike sampling limit, changed observer, or an enlarged trial class
Fermionic HadamardDirac bisolutions with CAR complementarity, 0≤C±≤10\le C^\pm\le1, and spin structureGenerally nonunique; zero-mode occupation may remainOptional stationary or geometric inputBundle-valued Hadamard relationRequires the CAR version of the local theoremMissing spin structure, wrong first-order constraint, or failed CAR positivity
Gauge-field HadamardGreen-hyperbolic gauge complex, constraints or BRST quotient, physical commutator, and physical positivityGauge representatives are nonunique; physical sectors can also differOptional gauge-invariant geometric inputBundle-valued Hadamard control on appropriate representatives or observablesTopology, boundary conditions, and superselection sector must be fixedTreating an indefinite potential or ghost kernel as a positive physical covariance
Instantaneous proposalHamiltonian diagonalization on one slice in chosen canonical variablesRelative only to that slice and parametrizationInstantaneous time choiceOften insufficient unless upgraded by controlled high-order dataNo general Hadamard local-comparison conclusionRapid background variation, negative instantaneous frequency squared, or slice dependence
de Sitter alpha proposalNormalized antipodal Bogoliubov mixing of the Euclidean stateA nontrivial invariant familyde Sitter invariantNon-Hadamard except at the Euclidean memberNot generally in the Hadamard local foliumOpposite-frequency and antipodal singularities
Sorkin–Johnston proposalPositive spectral part of iΔ=−iEi\Delta=-iE on a specified region and operator domainFixed relative to that region and spectral problemInherits symmetries preserved by the regionGenerally non-HadamardNot generally locally quasiequivalent to a Hadamard stateRegion, boundary, zero-mode, or continuum-limit dependence

Continue to Interacting Fields and Local Covariant Renormalization to understand why ultraviolet subtraction must respect local geometry.

  • Radzikowski, Marek J. “Micro-Local Approach to the Hadamard Condition in Quantum Field Theory on Curved Space-Time.” Communications in Mathematical Physics 179 (1996): 529–553. DOI.