Energy–Time Uncertainty

Not quite a Heisenberg relation

You have already seen the position–momentum uncertainty principle,

$\Delta x \,\Delta p \;\geq\; \frac{\hbar}{2},$

which places a simultaneous lower bound on two complementary observables. A superficially similar statement appears throughout quantum physics:

$\Delta E \,\Delta t \;\geq\; \frac{\hbar}{2}.$

The notation looks identical, but the meaning is fundamentally different — and the difference matters for correctly interpreting experimental results.

In the position–momentum relation, $\Delta x$ and $\Delta p$ are standard deviations of two quantum-mechanical observables measured on the same state. Time, by contrast, is not an observable in ordinary quantum mechanics; it is an external parameter that labels when a measurement is performed. There is no Hermitian operator $\hat{t}$ whose eigenstates could be prepared or whose expectation value could be minimised.

What $\Delta t$ actually measures

The most common and experimentally useful interpretation of $\Delta t$ is the characteristic timescale over which the quantum state changes appreciably. More precisely, Mandelstam and Tamm (1945) proved a precise theorem: for any observable $\hat{Q}$,

$\frac{\mathrm{d}\langle \hat{Q} \rangle}{\mathrm{d}t} = \frac{i}{\hbar}\langle [\hat{H}, \hat{Q}] \rangle,$

which, combined with the Robertson relation for $\hat{H}$ and $\hat{Q}$, gives

$\Delta E \,\cdot\, \frac{\Delta Q}{|\mathrm{d}\langle \hat{Q}\rangle / \mathrm{d}t|} \;\geq\; \frac{\hbar}{2}.$

The ratio $\Delta t_Q \equiv \Delta Q \,/\, |\mathrm{d}\langle \hat{Q}\rangle / \mathrm{d}t|$ is the time it takes $\langle \hat{Q} \rangle$ to change by one standard deviation $\Delta Q$ — a natural "lifetime" for the observable. The resulting inequality

$\Delta E \,\Delta t_Q \;\geq\; \frac{\hbar}{2}$

now holds as a rigorous theorem. Here $\Delta E = \Delta H$ is the spread in the energy of the state, and $\Delta t_Q$ depends on which observable you track — there is no single, universal clock attached to a quantum state.

Spectral linewidths: a concrete consequence

The most direct physical application is the natural linewidth of an atomic transition. An excited atomic state has a finite mean lifetime $\tau$ — the average time before it decays by emitting a photon. Because the state only exists for a finite duration, the energy it radiates cannot be perfectly sharp; the emitted photon has a frequency spread

$\Delta \nu \;\simeq\; \frac{1}{2\pi\tau},$

derived directly from $\Delta E \,\tau \sim \hbar$ and $\Delta E = h \,\Delta\nu$.

For the hydrogen $2p \to 1s$ transition, the excited state lifetime is approximately $\tau \approx 1.6 \times 10^{-9}\,\text{s}$, giving

$\Delta\nu \;\simeq\; \frac{1}{2\pi \times 1.6 \times 10^{-9}\,\text{s}} \;\approx\; 10^{8}\,\text{Hz}.$

The emitted photon's energy is about $10.2\,\text{eV} \approx 2.5 \times 10^{15}\,\text{Hz}$, so the fractional linewidth is roughly $\Delta\nu/\nu \sim 4 \times 10^{-8}$ — extremely narrow, but nonzero and, importantly, set by a fundamental quantum constraint rather than by instrumental imperfection.

Virtual processes and particle decay

In particle physics the same logic applies to unstable particles. A resonance with mean lifetime $\tau$ has an intrinsic energy width

$\Gamma \;\equiv\; \frac{\hbar}{\tau},$

called its decay width. The $Z^0$ boson, for example, has $\tau \approx 2.6 \times 10^{-25}\,\text{s}$, so $\Gamma \approx \hbar / (2.6 \times 10^{-25}) \approx 2.5\,\text{GeV}$ — in good agreement with the measured value of $2.495\,\text{GeV}$. The width is directly visible as the Breit–Wigner peak in $e^+e^-$ collision cross-sections.

The time–energy relation does not permit "borrowing" energy

A common but misleading statement is that the uncertainty principle "allows" a system to violate conservation of energy for a short time $\Delta t \sim \hbar / \Delta E$. This framing is not supported by modern quantum mechanics or quantum field theory. Energy is conserved in every interaction; what the time–energy relation governs is the resolvability of energy states and the spread of energies in states with finite lifetimes, not a transient suspension of conservation laws.

Summary

| Symbol | Meaning | |--------|---------| | $\Delta E$ | Standard deviation of energy in the quantum state | | $\Delta t$ | Characteristic time for the state to change (context-dependent) | | $\tau$ | Mean lifetime of an excited state or unstable particle | | $\Gamma = \hbar/\tau$ | Decay width (energy spread) |

The energy–time relation $\Delta E \,\Delta t \geq \hbar/2$ is a rigorous inequality when $\Delta t$ is interpreted as the Mandelstam–Tamm timescale. Its most direct observable consequence is the finite natural linewidth of spectral lines and the measurable decay widths of unstable particles — both proportional to $\hbar / \tau$.

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