|q⟩ Bad Qubits

beginner · Physics · The Uncertainty Principle

The Commutator Bound

The commutator

Two observables AA and BB do not, in general, commute. Their commutator is the operator

[A,B]ABBA.[A, B] \equiv AB - BA.

If [A,B]=0[A, B] = 0 the operators share a complete set of eigenstates and can in principle be measured simultaneously with arbitrary precision. When [A,B]0[A, B] \neq 0, a fundamental constraint kicks in.

Robertson's inequality

In 1929, H. P. Robertson derived a general lower bound on the product of standard deviations of any two observables. For a system in state ψ|\psi\rangle, let σA\sigma_A and σB\sigma_B be the standard deviations of AA and BB respectively. Then

σAσB    12[A,B],\sigma_A \,\sigma_B \;\geq\; \frac{1}{2}\,\bigl|\langle [A, B] \rangle\bigr|,

where [A,B]=ψ[A,B]ψ\langle [A, B] \rangle = \langle \psi | [A, B] | \psi \rangle is the expectation value of the commutator in that state.

The inequality follows from the Cauchy–Schwarz inequality applied to the vectors f=(AA)ψ|f\rangle = (A - \langle A\rangle)|\psi\rangle and g=(BB)ψ|g\rangle = (B - \langle B\rangle)|\psi\rangle. One writes fg=Refg+iImfg\langle f | g \rangle = \text{Re}\langle f|g\rangle + i\,\text{Im}\langle f|g\rangle and notes that the imaginary part equals 12i[A,B]\frac{1}{2i}\langle [A,B]\rangle (so that Imfg=12[A,B]|\text{Im}\langle f|g\rangle| = \frac{1}{2}|\langle [A,B]\rangle|, recalling that [A,B]\langle[A,B]\rangle is purely imaginary for Hermitian A,BA,B), so the Cauchy–Schwarz bound f2g2fg2Imfg2\|f\|^2\|g\|^2 \geq |\langle f|g\rangle|^2 \geq |\text{Im}\langle f|g\rangle|^2 gives exactly Robertson's result.

Position and momentum

The canonical commutation relation,

[x,p]=i,[x, p] = i\hbar,

is the foundational algebraic fact of non-relativistic quantum mechanics. Because [x,p][x,p] equals ii\hbar times the identity, the expectation value [x,p]=i\langle [x, p]\rangle = i\hbar in every normalizable state, and Robertson's inequality immediately gives

σxσp    12i=2.\sigma_x \,\sigma_p \;\geq\; \frac{1}{2}\,|i\hbar| = \frac{\hbar}{2}.

This is the Heisenberg uncertainty principle, now understood as a theorem about operators rather than a statement about experimental disturbance. The numerical bound is

2=1.0545718×1034J⋅s25.27×1035J⋅s.\frac{\hbar}{2} = \frac{1.0545718 \times 10^{-34}\,\text{J·s}}{2} \approx 5.27 \times 10^{-35}\,\text{J·s}.

States that saturate the bound — Gaussian wave packets for which σxσp=/2\sigma_x \sigma_p = \hbar/2 — are called minimum-uncertainty states. The harmonic-oscillator coherent states are the canonical example (they keep equal, balanced spreads in xx and pp); the squeezed states are minimum-uncertainty packets too, trading a smaller spread in one variable for a larger spread in the other. All non-Gaussian states have a strictly larger product.

Other pairs

The Robertson inequality applies to any two observables. For the components of angular momentum, [Lx,Ly]=iLz[L_x, L_y] = i\hbar L_z, so

σLxσLy    2Lz.\sigma_{L_x}\,\sigma_{L_y} \;\geq\; \frac{\hbar}{2}\,|\langle L_z \rangle|.

Here the bound does depend on the state through Lz\langle L_z \rangle. In an eigenstate of LzL_z with quantum number m=0m=0 the right-hand side vanishes, so Robertson's inequality imposes no constraint at all. A vanishing bound does not imply the two observables become sharp, however: for any 1\ell \geq 1 that same m=0m=0 state has σLx=σLy=(+1)/2>0\sigma_{L_x} = \sigma_{L_y} = \hbar\sqrt{\ell(\ell+1)/2} > 0. The inequality is simply silent there.

Try it

This is a numerical exercise — return a number, not a circuit. Using =1.0545718×1034J⋅s\hbar = 1.0545718 \times 10^{-34}\,\text{J·s}, compute the lower bound on σxσp\sigma_x \,\sigma_p implied by the Robertson inequality and the canonical commutation relation.

Run your code to see the quantum state.

Sign in on the full site to ask questions and join the discussion.