{
  "id": "fermis-golden-rule",
  "title": "Fermi's Golden Rule",
  "lang": "en",
  "source_lang": "en",
  "summary": "Fermi’s golden rule gives the transition rate from a quantum state into a continuum of final states under a weak perturbation.",
  "category": "science",
  "subcategory": "physics",
  "type": "concept",
  "aliases": [
    "Fermi golden rule",
    "Golden rule of quantum transitions"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/fermis-golden-rule/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/fermis-golden-rule/",
      "json": "https://aiwiki.page/zh/fermis-golden-rule.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Fermi’s golden rule is an approximate formula in quantum mechanics for the rate at which a weak interaction causes transitions from an initial quantum state into a continuum, or sufficiently dense set, of final states. It relates the transition rate to the squared interaction matrix element and the number of accessible final states per unit energy. The rule is a central result of time-dependent perturbation theory and is widely used to calculate radiative transitions, particle decays, scattering rates, and relaxation processes in solids."
    },
    {
      "heading": "Basic formulation",
      "text": "Consider a system with Hamiltonian\n\n\\[H=H_0+V,\\]\n\nwhere \\(H_0\\) describes the unperturbed system and \\(V\\) is a weak interaction. Let the initial state \\(|i\\rangle\\) have unperturbed energy \\(E_i\\), and let \\(|f\\rangle\\) denote an accessible final state with energy \\(E_f\\). The transition rate is conventionally written\n\n\\[\\Gamma_{i\\rightarrow f} = \\frac{2\\pi}{\\hbar} \\left|\\langle f|V|i\\rangle\\right|^2 \\delta(E_f-E_i),\\]\n\nwhere \\(\\hbar=h/(2\\pi)\\) is the reduced Planck constant and \\(\\delta\\) is the Dirac delta distribution.\n\nThis expression is a distribution to be summed or integrated over final states, rather than an ordinary finite rate into one isolated discrete state. The total transition rate is\n\n\\[\\Gamma_i = \\frac{2\\pi}{\\hbar} \\sum_f \\left|\\langle f|V|i\\rangle\\right|^2 \\delta(E_f-E_i),\\]\n\nwith a continuum integral replacing the sum when appropriate.\n\nIf the final states have a density of states \\(\\rho(E)\\), and the matrix element has a common value or a suitable average on the relevant energy shell, the familiar form is\n\n\\[\\boxed{ \\Gamma_i = \\frac{2\\pi}{\\hbar} |V_{fi}|^2\\rho(E_i) }.\\]\n\nHere:\n\n- \\(V_{fi}=\\langle f|V|i\\rangle\\) is the interaction matrix element.\n- \\(\\rho(E)\\,dE\\) is the number of accessible final states in an energy interval \\(dE\\).\n- \\(\\Gamma_i\\) is a transition probability per unit time, not a probability itself.\n\nFor normalized discrete states, \\(V_{fi}\\) has units of energy and \\(\\rho(E)\\) has units of inverse energy, so the rate has units of inverse time. Other normalization conventions redistribute dimensions between the matrix element and the state-counting measure without changing the physical result."
    },
    {
      "heading": "Derivation from time-dependent perturbation theory",
      "text": "Suppose the system is initially in \\(|i\\rangle\\), and a time-independent perturbation is switched on at \\(t=0\\). To first order in \\(V\\), the transition amplitude for \\(f\\ne i\\) is\n\n\\[c_f^{(1)}(t) = -\\frac{i}{\\hbar}V_{fi} \\int_0^t e^{i(E_f-E_i)t'/\\hbar}\\,dt'.\\]\n\nThis follows by expanding the Schrödinger equation in the unperturbed energy eigenstates. Applying the Born rule gives\n\n\\[P_{i\\rightarrow f}(t) = 4|V_{fi}|^2 \\frac{ \\sin^2[(E_f-E_i)t/(2\\hbar)] }{ (E_f-E_i)^2 }.\\]\n\nAt exact resonance, its limiting value is\n\n\\[P_{i\\rightarrow f}(t) = \\frac{|V_{fi}|^2t^2}{\\hbar^2}.\\]\n\nThus, first-order perturbation theory does not give a constant irreversible rate between two isolated states. The golden-rule rate emerges only after summing over a sufficiently dense set of final states.\n\nThe sharply peaked factor in the transition probability obeys the distributional limit\n\n\\[\\lim_{t\\rightarrow\\infty} \\frac{1}{t} \\frac{ 4\\sin^2(\\Delta E\\,t/2\\hbar) }{ (\\Delta E)^2 } = \\frac{2\\pi}{\\hbar}\\delta(\\Delta E).\\]\n\nIts width in energy decreases approximately as \\(\\hbar/t\\), while its integrated area grows linearly with time. Integrating against a smooth density of final states therefore gives\n\n\\[P_{\\mathrm{out}}(t)\\approx \\Gamma_i t.\\]\n\nThe infinite-time limit is a mathematical device for identifying the rate. Physically, the derivation applies over an intermediate time interval: long enough to resolve the relevant continuum, but short enough that depletion of the initial state remains negligible in the first-order calculation."
    },
    {
      "heading": "Energy conservation and periodic driving",
      "text": "For a time-independent total Hamiltonian, the delta distribution selects final states with \\(E_f=E_i\\). This expresses conservation of total energy. A subsystem may nevertheless gain or lose energy if another part of the full system gains or loses a compensating amount.\n\nFor a periodic perturbation,\n\n\\[V(t)=W e^{-i\\omega t}+W^\\dagger e^{i\\omega t},\\]\n\nthe two components give\n\n\\[\\Gamma_{i\\rightarrow f} = \\frac{2\\pi}{\\hbar} \\left[ |\\langle f|W|i\\rangle|^2 \\delta(E_f-E_i-\\hbar\\omega) + |\\langle f|W^\\dagger|i\\rangle|^2 \\delta(E_f-E_i+\\hbar\\omega) \\right].\\]\n\nThe first term describes the system gaining energy \\(\\hbar\\omega\\); the second describes it losing that amount. In an interaction with a quantized electromagnetic field, these processes correspond to absorption and emission of a photon.\n\nThe amplitude convention matters. If a perturbation is written as \\(A\\cos\\omega t\\), each complex exponential has amplitude \\(A/2\\), which contributes a factor of \\(1/4\\) when squared."
    },
    {
      "heading": "Matrix elements, state counting, and selection rules",
      "text": "The two principal ingredients of the rule describe different physical constraints.\n\nInteraction matrix element\n\nThe matrix element measures how strongly a particular interaction connects the initial and final states. It depends on their wave functions, the form of the interaction, and the system’s symmetries.\n\nFor example, in the electric-dipole approximation, the interaction with an applied electric field is\n\n\\[V(t)=-\\mathbf d\\cdot\\mathbf E(t),\\]\n\nwhere \\(\\mathbf d\\) is the electric dipole operator. The transition strength depends on\n\n\\[|\\langle f|\\mathbf d\\cdot\\boldsymbol{\\epsilon}|i\\rangle|^2,\\]\n\nwhere \\(\\boldsymbol{\\epsilon}\\) specifies the field’s polarization.\n\nSymmetry may force this matrix element to vanish, producing a selection rule. In atoms, electric-dipole selection rules constrain changes in angular momentum and parity. A transition forbidden in the electric-dipole approximation can still occur through weaker interactions, such as magnetic-dipole or electric-quadrupole coupling.\n\nDensity and occupation of final states\n\nThe density of states determines how many final states are available at the required energy. It is not the same as the ordinary mass density or particle density of a material.\n\nIf several final-state channels exist, their rates must be summed with their respective matrix elements and state-counting factors. A large density of states does not ensure a large transition rate if the corresponding matrix elements are small or zero.\n\nIn many-particle systems, occupation also matters. For fermions, the Pauli exclusion principle blocks transitions into already occupied single-particle states. In a noninteracting or quasiparticle description, transition rates commonly contain a factor\n\n\\[f_i(1-f_f),\\]\n\nwhere \\(f_i\\) and \\(f_f\\) are initial- and final-state occupation probabilities. For bosons, occupation of the final mode can instead enhance a transition through a factor \\(1+n_f\\)."
    },
    {
      "heading": "Applications",
      "text": "Optical transitions and spontaneous emission\n\nThe rule connects microscopic light–matter coupling to absorption and emission rates in spectroscopy. For spontaneous emission, the final states include both the lower-energy matter state and an emitted photon. The electromagnetic vacuum supplies the available photon modes; an initially populated radiation field is not required.\n\nBecause the rate depends on the available electromagnetic modes, changing the surrounding optical environment can alter spontaneous-emission rates.\n\nScattering and transport\n\nIn scattering theory, the rule gives transition rates between incoming and outgoing states. Dividing a suitable rate by incident flux produces a scattering cross section.\n\nIn condensed-matter physics, it is used to calculate electron scattering by impurities, other electrons, and phonons. These microscopic rates enter transport models for conductivity, mobility, and energy relaxation. A transport relaxation time need not equal the total scattering time, because different scattering angles relax momentum by different amounts.\n\nNuclear and particle transitions\n\nThe rule underlies perturbative calculations of decay rates in nuclear physics and particle physics. In beta decay, for example, both the weak-interaction matrix element and the phase space available to the outgoing particles determine the decay rate and energy spectrum.\n\nRelaxation and population kinetics\n\nGolden-rule rates are often used as inputs to equations describing populations of quantum states. In this role, the rule connects microscopic interactions to approximately irreversible kinetics. The rate equation itself requires additional assumptions about correlations, coherence, and the persistence of memory in the environment."
    },
    {
      "heading": "Validity and limitations",
      "text": "The golden rule is not an exact law for arbitrary quantum dynamics.\n\nWeak coupling. The leading nonzero perturbative contribution must dominate. The standard formula uses a first-order transition amplitude and is therefore second order in the interaction strength. Higher-order processes can sometimes be treated using an effective transition matrix element.\n\nA continuum or effectively dense spectrum. A few isolated, coherently coupled states generally exhibit oscillatory population transfer rather than irreversible transitions. A resonantly driven two-level system, for example, can undergo Rabi oscillations.\n\nA suitable time window. The observation time must be long enough that the finite-time energy peak is narrow compared with variations in the coupling and state density. It must also remain short compared with significant initial-state depletion in the elementary derivation. In a finite system, recurrences can eventually invalidate a constant-rate description.\n\nSmooth spectral structure. Near thresholds, band edges, sharp resonances, or singularities in the density of states, replacing the full finite-time expression by an energy-conserving delta distribution can fail. Strong environmental memory can likewise prevent a constant-rate approximation.\n\nSeparation from exponential decay. The direct perturbative result is\n\n\\[P_{\\mathrm{out}}(t)\\approx\\Gamma t.\\]\n\nAn exponential survival probability,\n\n\\[P_i(t)=e^{-\\Gamma t},\\]\n\nfollows when an approximately constant instantaneous decay rate remains applicable as the initial population decreases. It is not the direct first-order result at all times. Exact survival probabilities generally have quadratic short-time behavior and can show nonexponential long-time behavior.\n\nFor an isolated resonance with approximately exponential population decay, the lifetime is \\(\\tau=1/\\Gamma\\), and the associated Lorentzian energy width is\n\n\\[\\Delta E=\\hbar\\Gamma=\\frac{\\hbar}{\\tau}.\\]\n\nHere \\(\\Delta E\\) denotes the full width at half maximum. Additional dephasing or broadening mechanisms can modify an observed spectral linewidth."
    },
    {
      "heading": "Historical development",
      "text": "The rule is named after Enrico Fermi, who emphasized its usefulness and helped establish the terminology “golden rule.” Its foundations belong to the early development of time-dependent perturbation theory. Paul Dirac had already used closely related transition-rate methods in his 1927 treatment of the emission and absorption of radiation.\n\nThe name therefore identifies a widely used result associated with Fermi’s exposition, rather than a result whose entire mathematical development originated with him alone."
    }
  ],
  "links": [
    "quantum-mechanics",
    "perturbation-theory",
    "hamiltonian-operator",
    "planck-constant",
    "density-of-states",
    "schrodinger-equation",
    "born-rule",
    "energy",
    "photon",
    "wave-function",
    "electric-dipole",
    "polarization",
    "selection-rules",
    "angular-momentum",
    "fermion",
    "pauli-exclusion-principle",
    "boson",
    "spectroscopy",
    "scattering",
    "condensed-matter-physics",
    "phonon",
    "nuclear-physics",
    "beta-decay",
    "enrico-fermi"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}