Lesson 259: Operators as Matrices in Hilbert Space

Introduction: What Operators Do

An operator \(\hat{A}\) is a function that maps states to states: \(\hat{A}|\psi\rangle = |\psi'\rangle\). In a finite-dimensional Hilbert space with a chosen basis, every operator can be represented as a matrix. In infinite dimensions, operators are more subtle but the matrix analogy still guides our intuition.

Matrix Elements

Given an orthonormal basis \(\{|n\rangle\}\), the matrix elements of operator \(\hat{A}\) are:

\[A_{mn} = \langle m|\hat{A}|n\rangle\]

The matrix representation is:

\[A = \begin{pmatrix} \langle 1|\hat{A}|1\rangle & \langle 1|\hat{A}|2\rangle & \cdots \\ \langle 2|\hat{A}|1\rangle & \langle 2|\hat{A}|2\rangle & \cdots \\ \vdots & \vdots & \ddots \end{pmatrix}\]

Operator Algebra

Worked Examples

Example 1: Matrix of the Pauli X Operator

In the \(\{|0\rangle, |1\rangle\}\) basis, if \(\hat{\sigma}_x|0\rangle = |1\rangle\) and \(\hat{\sigma}_x|1\rangle = |0\rangle\):

\[\sigma_x = \begin{pmatrix} \langle 0|\hat{\sigma}_x|0\rangle & \langle 0|\hat{\sigma}_x|1\rangle \\ \langle 1|\hat{\sigma}_x|0\rangle & \langle 1|\hat{\sigma}_x|1\rangle \end{pmatrix} = \begin{pmatrix} 0 & 1 \\ 1 & 0 \end{pmatrix}\]

Example 2: Applying an Operator

Apply \(\hat{\sigma}_x\) to \(|\psi\rangle = \frac{1}{\sqrt{2}}(|0\rangle + |1\rangle)\):

\[\hat{\sigma}_x|\psi\rangle = \frac{1}{\sqrt{2}}(\hat{\sigma}_x|0\rangle + \hat{\sigma}_x|1\rangle) = \frac{1}{\sqrt{2}}(|1\rangle + |0\rangle) = |\psi\rangle\]

The state is unchanged—it's an eigenstate with eigenvalue 1!

Example 3: Computing a Matrix Element

For the lowering operator \(\hat{a}\) with \(\hat{a}|n\rangle = \sqrt{n}|n-1\rangle\):

\[\langle m|\hat{a}|n\rangle = \sqrt{n}\langle m|n-1\rangle = \sqrt{n}\delta_{m,n-1}\]

Non-zero only when \(m = n-1\).

The Quantum Connection

In quantum mechanics, physical quantities (position, momentum, energy) are represented by operators. The expectation value is:

\[\langle\hat{A}\rangle = \langle\psi|\hat{A}|\psi\rangle\]

This is the average value you'd measure over many identical experiments. The matrix representation makes calculation concrete: it's just matrix-vector multiplication.