logo

Hilbert Space's Orthonormal Basis and Unitary Operator 📂Hilbert Space

Hilbert Space's Orthonormal Basis and Unitary Operator

Definition

If a Schauder basis {ek}kN\left\{ \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} of a Hilbert space HH is a normal orthogonal system, then {ek}kN\left\{ \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} is called the Orthonormal Basis of HH.

Theorem1

Equivalent Conditions for Orthonormal Basis

  • [1]: Assuming HH is a Hilbert space. For the normal orthogonal system {ek}kNH\left\{ \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} \subset H of HH, the following are all equivalent.
    • (i): {ek}kNH\left\{ \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} \subset H is the Orthonormal Basis of HH.
    • (ii): For all xH\mathbf{x}\in H: x=kNx,ekek \mathbf{x}= \sum_{k \in \mathbb{N}} \langle \mathbf{x}, \mathbf{e}_{k} \rangle \mathbf{e}_{k}
    • (iii): For all x,yH\mathbf{x}, \mathbf{y} \in H: x,y=kNx,ekek,y \langle \mathbf{x}, \mathbf{y} \rangle = \sum_{k \in \mathbb{N}} \langle \mathbf{x}, \mathbf{e}_{k} \rangle \langle \mathbf{e}_{k} , \mathbf{y} \rangle
    • (iv): For all xH\mathbf{x}\in H: kNx,ek2=x2 \sum_{k \in \mathbb{N}} \left| \langle \mathbf{x}, \mathbf{e}_{k} \rangle \right|^{2} = \left\| \mathbf{x}\right\|^{2}
    • (v): span{ek}kN=H\overline{\text{span}} \left\{ \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} = H
    • (vi): If xH\mathbf{x}\in H and for all kNk \in \mathbb{N} if x,ek=0\langle \mathbf{x}, \mathbf{e}_{k} \rangle = 0 then x=0\mathbf{x}= \mathbf{0}

Unitary Operators and Orthonormal Basis

  • [2]: If {ek}kN\left\{ \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} is considered the orthonormal basis of HH, then the orthonormal basis of HH is exactly represented as {Uek}kN\left\{ U \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} with respect to the unitary operator U:HHU : H \to H.

Description

Especially, it is said that all orthonormal bases of HH are characterized by the unitary operator UU, similar to the result in theorem [2].

Proof

Refer to references for the proof of [1].

[2]

Let also {vk}kN\left\{ \mathbf{v}_{k} \right\}_{k \in \mathbb{N}} be the orthonormal basis of HH. Define the operator U:HHU : H \to H as follows: U(kNckek):=kNckvk,ckkNl2 U \left( \sum_{k \in \mathbb{N}} c_{k} \mathbf{e}_{k} \right) := \sum_{k \in \mathbb{N}} c_{k} \mathbf{v}_{k} \qquad , \forall {c_{k}}_{k \in \mathbb{N}} \in l^{2} Then, UU is bounded, bijective, and vk=Uek\mathbf{v}_{k} = U \mathbf{e}_{k}.

Since {ek}kN\left\{ \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} is the orthonormal basis of HH, according to (i)     \implies (ii), v,wH\mathbf{v} ,\mathbf{w} \in H can be represented as follows:

v=kNv,ekekw=kNw,ekek \mathbf{v} = \sum_{k \in \mathbb{N}} \left\langle \mathbf{v} , \mathbf{e}_{k} \right\rangle \mathbf{e}_{k} \\ \mathbf{w} = \sum_{k \in \mathbb{N}} \left\langle \mathbf{w} , \mathbf{e}_{k} \right\rangle \mathbf{e}_{k}

Then, by the definition of UU and (i)     \implies (iii),

UUv,w=Uv,Uw=kNv,ekek,kNw,ekek=kNv,ekw,ek=v,w \begin{align*} \left\langle U^{ \ast } U \mathbf{v} , \mathbf{w} \right\rangle =& \left\langle U \mathbf{v} , U \mathbf{w} \right\rangle \\ =& \left\langle \sum_{k \in \mathbb{N}} \left\langle \mathbf{v} , \mathbf{e}_{k} \right\rangle \mathbf{e}_{k} , \sum_{k \in \mathbb{N}} \left\langle \mathbf{w} , \mathbf{e}_{k} \right\rangle \mathbf{e}_{k} \right\rangle \\ =& \sum_{k \in \mathbb{N}} \left\langle \mathbf{v} , \mathbf{e}_{k} \right\rangle \overline{\left\langle \mathbf{w} , \mathbf{e}_{k} \right\rangle} \\ =& \left\langle \mathbf{v} , \mathbf{w} \right\rangle \end{align*}

In other words, since UU=IU^{ \ast } U = I, UU is a unitary operator, and it is bijective having the inverse operator U1=UU^{-1} = U^{ \ast }. Meanwhile, assuming UU is unitary from the assumption produces

Uei,Uej=UUei,ej=ei,ej=δij \left\langle U \mathbf{e}_{i} , U \mathbf{e}_{j} \right\rangle = \left\langle U^{ \ast } U \mathbf{e}_{i} , \mathbf{e}_{j} \right\rangle = \left\langle \mathbf{e}_{i} , \mathbf{e}_{j} \right\rangle = \delta_{ij}

That is, {Uek}kN\left\{ U \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} is an orthonormal set. To show that it becomes the basis of HH, let’s assume v,Uek=0\left\langle \mathbf{v} , U \mathbf{e}_{k} \right\rangle = 0 for all kNk \in \mathbb{N}. Then, for all kNk \in \mathbb{N}, Uv,ek=0\left\langle U^{ \ast } \mathbf{v} , \mathbf{e}_{k} \right\rangle = 0, thus Uv=0U^{ \ast } \mathbf{v} = \mathbf{0} must hold. It was already shown that U=U1U^{ \ast } = U^{-1}, so applying UU to both sides yields v=0\mathbf{v} = \mathbf{0}. Consequently, according to (vi)     \implies (i), it can be confirmed that {Uek}kNH\left\{ U \mathbf{e}_{k} \right\}_{k \in \mathbb{N}} \subset H becomes the orthonormal basis of HH.


  1. Ole Christensen, Functions, Spaces, and Expansions: Mathematical Tools in Physics and Engineering (2010), p80-83 ↩︎