Fermi's Golden Rule
So far transitions went to a single discrete final state, giving a probability that oscillates or grows quadratically. In most real situations — an atom emitting into the continuum of photon modes, an electron scattering into a band of momentum states — the final state belongs to a continuum. Summing over that continuum converts the awkward growth into a clean, constant rate. The result is Fermi's golden rule.
Summing over a continuum of final states
The total probability of leaving is obtained by summing over all accessible final states. For a continuum we replace the sum by an integral weighted by the density of states — the number of final states per unit energy:
using the constant-perturbation result with .
The sinc-squared becomes a delta function
The factor is the narrowing peak we met before. For large it behaves like a (scaled) Dirac delta in energy. Precisely,
which, written in energy with , gives . Substituting and assuming and vary slowly across the narrow peak, the integral collapses:
A constant rate
Because is now linear in , the transition rate is constant:
evaluated at the final energy fixed by energy conservation. This is Fermi's golden rule. For a harmonic perturbation of amplitude the same derivation gives with the resonance condition .
Reading the ingredients
- — how strongly the perturbation connects initial and final states. Vanishing matrix elements give selection rules (zero rate).
- — how many final states are available at the right energy. More available states means a faster transition. This is why decay into the open continuum is irreversible while transitions between two isolated levels merely oscillate.
- (or with a drive) — energy conservation, enforced by the delta function.
Try it
Apply the golden rule. With coupling , density of states , and
, compute the transition rate and return
it. (Remember to square the matrix element.)
Sign in on the full site to ask questions and join the discussion.