Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5.7 (RAGE) — dynamical characterization of ℌ_c and ℌ_pp

Proved
TeschlQM.Dynamics.rage

by mikedeng1 · Sep 29, 2026 · Mathlib 0df444a (Lean v4.33.1)

functional-analysisp2o-batch-books5p2o-gran-per-chapterp2o-plan-bookp2o-v1quantum-dynamicsspectral-theory

Let AAA be self-adjoint in a complex Hilbert space H\mathfrak HH. Suppose Kn∈L(H)K_n \in \mathfrak L(\mathfrak H)Kn​∈L(H) is a sequence of relatively compact operators which converges strongly to the identity. Then

Hc={ψ∈H  ∣  lim⁡n→∞lim⁡T→∞1T∫0T∥Kne−itAψ∥ dt=0},\mathfrak H_c = \Big\{\psi \in \mathfrak H \;\Big|\; \lim_{n\to\infty} \lim_{T\to\infty} \frac1T \int_0^T \|K_n \mathrm e^{-\mathrm itA}\psi\|\, dt = 0\Big\},Hc​={ψ∈H​n→∞lim​T→∞lim​T1​∫0T​∥Kn​e−itAψ∥dt=0}, Hpp={ψ∈H  ∣  lim⁡n→∞sup⁡t≥0∥(I−Kn)e−itAψ∥=0}.\mathfrak H_{pp} = \Big\{\psi \in \mathfrak H \;\Big|\; \lim_{n\to\infty} \sup_{t\ge 0} \|(\mathbb I - K_n)\mathrm e^{-\mathrm itA}\psi\| = 0\Big\}.Hpp​={ψ∈H​n→∞lim​t≥0sup​∥(I−Kn​)e−itAψ∥=0}.

Here Hc=Hac⊕Hsc\mathfrak H_c = \mathfrak H_{ac} \oplus \mathfrak H_{sc}Hc​=Hac​⊕Hsc​ and Hpp\mathfrak H_{pp}Hpp​ are the spectral subspaces defined from the spectral measures μψ\mu_\psiμψ​. The theorem (Ruelle, Amrein–Georgescu, Enß) says that continuous-spectrum states escape, on time average, from every "compact region" KnK_nKn​, while pure-point states stay uniformly localized for all times.

Formalization Note. e−itA\mathrm e^{-\mathrm itA}e−itA and the spectral subspaces are defined from a projection-valued measure PPP taken as data with A=∫λ dP(λ)A = \int\lambda\, dP(\lambda)A=∫λdP(λ); Hc\mathfrak H_cHc​ and Hpp\mathfrak H_{pp}Hpp​ are defined from the spectral measures, never from the dynamical conditions. The iterated limit means: for each nnn the limit LnL_nLn​ as T→∞T \to \inftyT→∞ exists, and Ln→0L_n \to 0Ln​→0. The supremum over t≥0t \ge 0t≥0 is taken in [0,∞][0,\infty][0,∞] (extended nonnegative reals), so it is never truncated. Separability of H\mathfrak HH is not assumed.

Preamble
import Mathlib
import Definitions.Def_TeschlQM_Shared_IsProjValuedMeasure
import Definitions.Def_TeschlQM_Shared_IsSpectralIntegral
import Definitions.Def_TeschlQM_Shared_timeEvolution
import Definitions.Def_TeschlQM_Dynamics_spectralSubspaces
import Definitions.Def_TeschlQM_Dynamics_IsRelativelyCompact

open Filter Topology
open scoped ENNReal
Formal statement
namespace TeschlQM.Dynamics

/-- Teschl, Theorem 5.7 (RAGE), p. 129, (5.14): let `A = ∫ λ dP(λ)` be self-adjoint and let
`K_n ∈ 𝔏(ℌ)` be relatively compact operators converging strongly to `𝕀`. Then
`ℌ_c = {ψ | lim_{n→∞} lim_{T→∞} (1/T) ∫₀ᵀ ‖K_n e^{−itA} ψ‖ dt = 0}` (for each `n` the inner limit
exists, and these limits tend to `0`) and
`ℌ_pp = {ψ | lim_{n→∞} sup_{t≥0} ‖(𝕀 − K_n) e^{−itA} ψ‖ = 0}` (the supremum taken in `[0, ∞]`).
`ℌ_c = ℌ_ac ⊕ ℌ_sc` and `ℌ_pp` are defined from the spectral measures `μ_ψ` (3.83). -/
theorem rage {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℂ H]
    [CompleteSpace H] (A : H →ₗ.[ℂ] H) (hA : IsSelfAdjoint A)
    (P : Set ℝ → (H →L[ℂ] H)) (hP : TeschlQM.Shared.IsProjValuedMeasure P)
    (hAP : TeschlQM.Shared.IsSpectralIntegral P (fun x : ℝ => (x : ℂ)) A)
    (K : ℕ → H →L[ℂ] H) (hK : ∀ n, IsRelativelyCompact ((K n : H →ₗ[ℂ] H).toPMap ⊤) A)
    (hKI : ∀ ψ : H, Tendsto (fun n => K n ψ) atTop (𝓝 ψ)) :
    hc P = {ψ : H | ∃ L : ℕ → ℝ,
      (∀ n, Tendsto (fun T : ℝ => (1 / T) * ∫ t in (0 : ℝ)..T, ‖K n (TeschlQM.Shared.timeEvolution P t ψ)‖)
        atTop (𝓝 (L n))) ∧ Tendsto L atTop (𝓝 0)} ∧
    hpp P = {ψ : H | Tendsto (fun n : ℕ => ⨆ t : Set.Ici (0 : ℝ),
      ‖TeschlQM.Shared.timeEvolution P t ψ - K n (TeschlQM.Shared.timeEvolution P t ψ)‖ₑ) atTop (𝓝 0)} := by sorry

end TeschlQM.Dynamics
Source
Teschl, Mathematical Methods in Quantum Mechanics, AMS GSM 99, 2009, p. 129, Theorem 5.7, Eq. (5.14)
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Let HHH be a complex inner product space that is complete, i.e. a complex Hilbert space. HHH may be the zero space. The statement's data and hypotheses are:

  • The operator AAA. AAA is a C\mathbb{C}C-linear operator on HHH that may be unbounded. It is defined on a linear subspace dom(A)⊆H\mathrm{dom}(A) \subseteq Hdom(A)⊆H, which need not be all of HHH.
  • Self-adjointness of AAA. AAA is assumed to equal its own adjoint, A∗=AA^* = AA∗=A. This uses the library's adjoint for partially defined operators. By that library's convention, if dom(A)\mathrm{dom}(A)dom(A) is not dense, the adjoint is the zero operator defined on all of HHH.
  • The map PPP. PPP assigns a bounded linear operator P(Ω)P(\Omega)P(Ω) on HHH to every subset Ω⊆R\Omega \subseteq \mathbb{R}Ω⊆R, not only to Borel sets. PPP is assumed to satisfy the imported predicate IsProjValuedMeasure. Its definition is not shown here; by its name it states that PPP is a projection-valued measure.
  • AAA as a spectral integral. The triple (P, λ↦λ∈C, A)(P,\ \lambda \mapsto \lambda \in \mathbb{C},\ A)(P, λ↦λ∈C, A) is assumed to satisfy the imported predicate IsSpectralIntegral. Its definition is not shown; by its name it states that AAA is the spectral integral ∫λ dP(λ)\int \lambda \, dP(\lambda)∫λdP(λ) of the real identity function, viewed as complex-valued, against PPP.
  • The operators KnK_nKn​. (Kn)n∈N(K_n)_{n \in \mathbb{N}}(Kn​)n∈N​, with n=0,1,2,…n = 0, 1, 2, \dotsn=0,1,2,…, is a sequence of bounded linear operators on HHH. For each nnn, KnK_nKn​ is regarded as an operator defined on all of HHH. The pair (Kn,A)(K_n, A)(Kn​,A) is assumed to satisfy the imported predicate IsRelativelyCompact. Its definition is not shown; by its name it states that KnK_nKn​ is relatively compact with respect to AAA.
  • Strong convergence to the identity. For every ψ∈H\psi \in Hψ∈H, Knψ→ψK_n \psi \to \psiKn​ψ→ψ in the norm of HHH as n→∞n \to \inftyn→∞.

Write UtψU_t\psiUt​ψ for the imported function timeEvolution evaluated at PPP, a real time ttt and a vector ψ\psiψ. Its definition is not shown, and it depends on PPP only, not directly on AAA. Likewise, Hc(P)\mathcal{H}_c(P)Hc​(P) and Hpp(P)\mathcal{H}_{pp}(P)Hpp​(P) denote the imported subsets hc P and hpp P of HHH. Their definitions are not shown, and they also depend on PPP only.

The theorem asserts both of the following.

First, Hc(P)\mathcal{H}_c(P)Hc​(P) is exactly the set of ψ∈H\psi \in Hψ∈H for which there is a real sequence (Ln)n∈N(L_n)_{n \in \mathbb{N}}(Ln​)n∈N​ such that:

  • for every nnn,
1T∫0T∥Kn Utψ∥ dt  ⟶  Lnas T→+∞ through the reals;\frac{1}{T}\int_0^T \big\| K_n\, U_t \psi \big\| \, dt \;\longrightarrow\; L_n \quad \text{as } T \to +\infty \text{ through the reals;}T1​∫0T​​Kn​Ut​ψ​dt⟶Ln​as T→+∞ through the reals;
  • and Ln→0L_n \to 0Ln​→0 as n→∞n \to \inftyn→∞.

The integral is an oriented interval integral of a real, nonnegative integrand. Membership therefore requires that each inner limit exists and is finite, and that these limits tend to 000.

Second, Hpp(P)\mathcal{H}_{pp}(P)Hpp​(P) is exactly the set of ψ∈H\psi \in Hψ∈H for which

sup⁡t∈[0,∞)∥Utψ−Kn Utψ∥  ⟶  0as n→∞.\sup_{t \in [0,\infty)} \big\| U_t\psi - K_n\, U_t \psi \big\| \;\longrightarrow\; 0 \quad \text{as } n \to \infty.t∈[0,∞)sup​​Ut​ψ−Kn​Ut​ψ​⟶0as n→∞.

Here the norm is taken as an extended nonnegative real, the supremum is taken in [0,∞][0,\infty][0,∞], and the convergence is in [0,∞][0,\infty][0,∞]. The supremum may equal +∞+\infty+∞ for some nnn. Convergence to 000 requires only that the suprema are eventually finite and tend to 000. Only the times t≥0t \ge 0t≥0 enter.

The operator AAA appears only in the hypotheses: self-adjointness, the spectral-integral relation with PPP, and relative compactness of each KnK_nKn​. It does not appear in either set equality.

Degenerate cases.

  • Integral and division conventions. If t↦∥KnUtψ∥t \mapsto \|K_n U_t\psi\|t↦∥Kn​Ut​ψ∥ is not integrable on [0,T][0,T][0,T], for example if it is not measurable, the integral is taken to be 000. For T=0T = 0T=0 the factor 1/T1/T1/T is taken to be 000. For T<0T < 0T<0 the integral is oriented from 000 to TTT. None of these affects a limit as T→+∞T \to +\inftyT→+∞ except through the first convention: if the integrand is non-integrable on [0,T][0,T][0,T] for all large TTT, the averages are eventually 000 and Ln=0L_n = 0Ln​=0.
  • Indexing. The sequence KnK_nKn​ starts at n=0n = 0n=0, and only its tail matters in either condition.
  • Nonempty supremum. The time set [0,∞)[0,\infty)[0,∞) is nonempty, so the supremum is never the supremum of an empty set.
  • Zero space HHH. Every vector is 000 and every KnK_nKn​ is 000, and both right-hand sets equal {0}=H\{0\} = H{0}=H. The statement then asserts that Hc(P)=Hpp(P)=H\mathcal{H}_c(P) = \mathcal{H}_{pp}(P) = HHc​(P)=Hpp​(P)=H.
  • Possibly vacuous hypotheses. Whether the hypotheses can all be satisfied at once depends on the bodies of the imported predicates IsProjValuedMeasure, IsSpectralIntegral and IsRelativelyCompact, which are not shown. If no data satisfy them, the theorem holds vacuously.
Human review
  • Endorsed by Shuze Chen · Oct 2, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 2, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me