Physlib

PhyslibAlpha.QuantumMechanics.OperatorAlgebra.Measurement.SpectralMeasure

The probability measure of an observable in a state

A state `ω` and an observable `a` together fix a probability measure `μ_{ω,a}` on `ℝ`: the one against which integrating `f` gives `ω(f(a))`, so it's what you integrate against for `a`'s expectation, variance, or any other statistic in the preparation `ω`. `μ_{ω,a}` is supported on `a`'s spectrum and is the unique measure with this property.

Restricting `Re ∘ ω` to the continuous functional calculus of `a` turns `ω, a` into one positive, normalized functional on `C(σ_ℝ(a), ℝ)` — exactly the input to Riesz–Markov–Kakutani, which Mathlib already supplies. Pushing the result forward along `σ_ℝ(a) ↪ ℝ` gives `μ_{ω,a}`.

14 declarations

definition

Spectral functional f↦Re(ω(f(a)))f \mapsto \text{Re}(\omega(f(a))) for a state ω\omega and observable aa

Given a C∗C^*-algebra AA, a state ω\omega on AA, and an observable a∈Aa \in A (a self-adjoint element), the spectral functional is the positive linear map from the space of continuous real-valued functions on the spectrum σ(a)\sigma(a), denoted C(σ(a),R)C(\sigma(a), \mathbb{R}), to R\mathbb{R}. For a function f∈C(σ(a),R)f \in C(\sigma(a), \mathbb{R}), the functional returns the real part of the state applied to the element f(a)f(a) obtained via the continuous functional calculus: f↦Re(ω(f(a)))f \mapsto \text{Re}(\omega(f(a))). Since ff is real-valued and aa is self-adjoint, f(a)f(a) is self-adjoint, making the result real-valued.

definition

Compactly supported spectral functional f↦Re(ω(f(a)))f \mapsto \text{Re}(\omega(f(a)))

Given a C∗C^*-algebra AA, a state ω\omega on AA, and an observable a∈Aa \in A (a self-adjoint element), this functional is the positive linear map from the space of continuous real-valued functions with compact support on the spectrum of aa, denoted Cc(σ(a),R)C_c(\sigma(a), \mathbb{R}), to R\mathbb{R}. For a function f∈Cc(σ(a),R)f \in C_c(\sigma(a), \mathbb{R}), the functional maps ff to Re(ω(f(a)))\text{Re}(\omega(f(a))), where f(a)f(a) is the element obtained via the continuous functional calculus. Since the spectrum σ(a)\sigma(a) is compact, this is the restriction of the spectral functional to the space of compactly supported functions, providing the necessary input for the Riesz–Markov–Kakutani representation theorem to produce a spectral measure.

definition

Spectral measure μω,a\mu_{\omega, a} on the spectrum σ(a)\sigma(a)

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and aa an observable (a self-adjoint element). The spectral measure associated with ω\omega and aa is the unique Borel measure μ\mu on the spectrum σ(a)⊆R\sigma(a) \subseteq \mathbb{R} such that for any continuous real-valued function f∈C(σ(a),R)f \in C(\sigma(a), \mathbb{R}), the integral of ff with respect to μ\mu satisfies: ∫σ(a)f dμ=Re(ω(f(a)))\int_{\sigma(a)} f \, d\mu = \text{Re}(\omega(f(a))) where f(a)f(a) is defined via the continuous functional calculus. This measure is obtained by applying the Riesz–Markov–Kakutani representation theorem to the positive linear functional f↦Re(ω(f(a)))f \mapsto \text{Re}(\omega(f(a))) on the space of continuous functions with compact support Cc(σ(a),R)C_c(\sigma(a), \mathbb{R}).

instance

The spectral measure μω,a\mu_{\omega, a} is finite

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and aa an observable (a self-adjoint element). The spectral measure μω,a\mu_{\omega, a} associated with ω\omega and aa—which is the unique Borel measure on the spectrum σ(a)⊆R\sigma(a) \subseteq \mathbb{R} such that ∫σ(a)f dμω,a=Re(ω(f(a)))\int_{\sigma(a)} f \, d\mu_{\omega, a} = \text{Re}(\omega(f(a))) for any continuous real-valued function ff—is a finite measure.

theorem

The spectral functional maps the constant function 1 to 1

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and aa an observable (a self-adjoint element). Let the spectral functional associated with ω\omega and aa be the linear map L:C(σ(a),R)→RL: C(\sigma(a), \mathbb{R}) \to \mathbb{R} defined by L(f)=Re(ω(f(a)))L(f) = \text{Re}(\omega(f(a))), where f(a)f(a) is defined by the continuous functional calculus. Then L(1)=1L(1) = 1, where 11 denotes the constant function on the spectrum σ(a)\sigma(a) that maps every element to 1.

instance

The spectral measure μω,a\mu_{\omega, a} is a probability measure

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and aa an observable (a self-adjoint element of AA). The spectral measure μω,a\mu_{\omega, a} associated with ω\omega and aa, which is a Borel measure defined on the spectrum σ(a)⊆R\sigma(a) \subseteq \mathbb{R}, is a probability measure. That is, its total mass is 1, which corresponds to the normalization property of the state ω(1)=1\omega(1) = 1.

theorem

ω(f(a))=∫f dμω,a\omega(f(a)) = \int f \, d\mu_{\omega, a}

Let AA be a unital C∗C^*-algebra, ω\omega a state on AA, and aa an observable (a self-adjoint element of AA). Let σ(a)⊆R\sigma(a) \subseteq \mathbb{R} denote the spectrum of aa, and let μω,a\mu_{\omega, a} be the spectral measure on σ(a)\sigma(a) associated with ω\omega and aa. For any continuous real-valued function f∈C(σ(a),R)f \in C(\sigma(a), \mathbb{R}), the state applied to the element f(a)f(a) (defined via the continuous functional calculus) satisfies: ω(f(a))=∫σ(a)f(x) dμω,a(x)\omega(f(a)) = \int_{\sigma(a)} f(x) \, d\mu_{\omega, a}(x) where the real-valued integral on the right is identified with its corresponding value in C\mathbb{C}.

definition

Spectral measure μω,a\mu_{\omega, a} on R\mathbb{R}

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and aa an observable (a self-adjoint element). The **real spectral measure** μω,a\mu_{\omega, a} is the probability measure on the real line R\mathbb{R} obtained by pushing forward the spectral measure (initially defined on the spectrum σ(a)⊆R\sigma(a) \subseteq \mathbb{R}) along the inclusion map ι:σ(a)↪R\iota: \sigma(a) \hookrightarrow \mathbb{R}. This measure is the unique probability measure on R\mathbb{R} supported on σ(a)\sigma(a) such that for any continuous real-valued function ff, the integral of ff with respect to μω,a\mu_{\omega, a} gives the expectation value: ∫Rf(x) dμω,a(x)=Re(ω(f(a)))\int_{\mathbb{R}} f(x) \, d\mu_{\omega, a}(x) = \text{Re}(\omega(f(a))) where f(a)f(a) is determined by the continuous functional calculus.

instance

μω,a\mu_{\omega, a} is a probability measure

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and a∈Aa \in A an observable (a self-adjoint element). The real spectral measure μω,a\mu_{\omega, a} associated with the state ω\omega and observable aa is a probability measure on R\mathbb{R}.

theorem

The spectral measure μω,a\mu_{\omega, a} is supported on the spectrum σ(a)\sigma(a)

Let AA be a C∗C^*-algebra, ω\omega be a state on AA, and aa be an observable. Let μω,a\mu_{\omega, a} denote the real spectral measure on R\mathbb{R} and σ(a)⊆R\sigma(a) \subseteq \mathbb{R} be the spectrum of aa. Then the measure of the complement of the spectrum σ(a)\sigma(a) with respect to μω,a\mu_{\omega, a} is zero: μω,a(σ(a)c)=0\mu_{\omega, a}(\sigma(a)^c) = 0 In other words, the probability measure μω,a\mu_{\omega, a} is concentrated on the spectrum of aa.

theorem

ω(f(a))=∫f dμω,a\omega(f(a)) = \int f \, d\mu_{\omega, a}

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and aa an observable (a self-adjoint element of AA). Let σ(a)⊆R\sigma(a) \subseteq \mathbb{R} denote the spectrum of aa and μω,a\mu_{\omega, a} denote the real spectral measure associated with the state ω\omega and observable aa. For any function f:R→Rf: \mathbb{R} \to \mathbb{R} that is continuous on the spectrum σ(a)\sigma(a), the expectation value of f(a)f(a) (determined by the continuous functional calculus) is given by the integral of ff with respect to μω,a\mu_{\omega, a}: ω(f(a))=∫Rf(x) dμω,a(x)\omega(f(a)) = \int_{\mathbb{R}} f(x) \, d\mu_{\omega, a}(x) where the integral on the right-hand side is a real number embedded into the complex numbers.

theorem

Extension of continuous functions on σ(a)\sigma(a) to R\mathbb{R}

Let aa be an observable (a self-adjoint element) in a C∗C^*-algebra AA, and let σ(a)⊆R\sigma(a) \subseteq \mathbb{R} denote its spectrum. For any continuous function g:σ(a)→Rg: \sigma(a) \to \mathbb{R} defined on the spectrum, there exists a function f:R→Rf: \mathbb{R} \to \mathbb{R} such that ff is continuous on σ(a)\sigma(a) and f(x)=g(x)f(x) = g(x) for all x∈σ(a)x \in \sigma(a).

theorem

Integral Equality for the Pullback of a Representative Measure and the Spectral Measure

Let AA be a C∗C^*-algebra, ω\omega a state on AA, and aa an observable. Let σ(a)⊆R\sigma(a) \subseteq \mathbb{R} denote the spectrum of aa and let ι:σ(a)→R\iota: \sigma(a) \to \mathbb{R} be the inclusion map. Suppose μ\mu is a measure on R\mathbb{R} such that: 1. For every function f:R→Rf: \mathbb{R} \to \mathbb{R} that is continuous on σ(a)\sigma(a), the expectation of the observable f(a)f(a) (defined via continuous functional calculus) satisfies ω(f(a))=∫Rf dμ\omega(f(a)) = \int_{\mathbb{R}} f \, d\mu. 2. The measure μ\mu is the pushforward of its pullback to the spectrum, i.e., μ=ι∗(ι∗μ)\mu = \iota_*(\iota^* \mu), which implies μ\mu is supported on σ(a)\sigma(a). Then, for any continuous function g∈C(σ(a),R)g \in C(\sigma(a), \mathbb{R}), the integral of gg with respect to the pulled-back measure ι∗μ\iota^* \mu is equal to the integral of gg with respect to the spectral measure μω,a\mu_{\omega, a}: ∫σ(a)g d(ι∗μ)=∫σ(a)g dμω,a\int_{\sigma(a)} g \, d(\iota^* \mu) = \int_{\sigma(a)} g \, d\mu_{\omega, a}

theorem

Uniqueness of the real spectral measure μω,a\mu_{\omega, a}

Let AA be a C∗C^*-algebra, ω\omega be a state on AA, and a∈Aa \in A be an observable (a self-adjoint element). Let σ(a)⊆R\sigma(a) \subseteq \mathbb{R} denote the spectrum of aa. Suppose μ\mu is a probability measure on R\mathbb{R} such that: 1. μ\mu is concentrated on the spectrum of aa, i.e., μ(R∖σ(a))=0\mu(\mathbb{R} \setminus \sigma(a)) = 0. 2. For every function f:R→Rf: \mathbb{R} \to \mathbb{R} that is continuous on σ(a)\sigma(a), the identity ω(f(a))=∫Rf(x) dμ(x)\omega(f(a)) = \int_{\mathbb{R}} f(x) \, d\mu(x) holds, where f(a)f(a) is defined via the continuous functional calculus. Then μ\mu is equal to the real spectral measure μω,a\mu_{\omega, a}.