Scattering theory
Scattering theory is the mathematical framework used to describe how waves or particles are redirected by interactions. Its central objects connect an asymptotic state prepared long before an interaction with another asymptotic state observed long afterward. Although the terminology suggests a physical projectile striking a localized target, the same structure applies to electromagnetic radiation, acoustic waves, condensed-matter excitations, and quantum fields.
A scattering process is characterized by measurable quantities such as the differential cross section, which specifies the distribution of outgoing flux with respect to direction or another final-state variable. The theory does not generally provide a detailed trajectory through the interaction region. Instead, it relates incoming and outgoing states through amplitudes whose absolute squares determine probabilities or flux ratios.
Asymptotic formulation
For a time-independent Hamiltonian, the dynamics are separated into a solvable free part (H_0) and an interaction (V),
[ H=H_0+V. ]
The free Hamiltonian defines the states that can be identified at large spatial separations or at asymptotically early and late times. If (|\phi\rangle) is a free state, the corresponding interacting in-state and out-state are written
[ |\psi^{(\pm)}\rangle
\Omega^{(\pm)}|\phi\rangle, ]
where (\Omega^{(\pm)}) are the Møller operators. Formally, these operators are defined by the strong limits
[ \Omega^{(\pm)}
\lim_{t\to\mp\infty} e^{iHt}e^{-iH_0t}, ]
when those limits exist on the scattering subspace. Bound states are not in the image of the Møller operators because they do not become freely separated states at asymptotic times.
The scattering matrix, conventionally denoted by (S), maps incoming free states to outgoing free states:
[ S=\Omega^{(-)\dagger}\Omega^{(+)}. ]
Conservation of total probability implies that (S) is unitary when all energetically accessible channels are included. This condition becomes nontrivial in multichannel problems because probability may leave the elastic channel while remaining within the complete Hilbert space. An apparently nonunitary elastic amplitude can therefore represent absorption into omitted reaction channels rather than a failure of quantum evolution.
For a continuous energy spectrum, the matrix is commonly decomposed as
[ S_{fi}
\delta_{fi}
2\pi i,\delta(E_f-E_i),T_{fi}, ]
where the transition matrix contains the dynamical information. The energy delta function expresses time-translation invariance and the resulting conservation of total energy.
Scattering states and the Lippmann–Schwinger equation
Stationary scattering states satisfy the Lippmann–Schwinger equation,
[ |\psi^{(\pm)}\rangle
|\phi\rangle + \frac{1}{E-H_0\pm i0},V|\psi^{(\pm)}\rangle. ]
The infinitesimal imaginary term specifies the boundary condition rather than an observable addition to the energy. The (+i0) prescription produces an outgoing scattered wave, while the (-i0) prescription produces the corresponding incoming solution. This distinction encodes the temporal orientation of the scattering state even though the underlying stationary equation contains no explicit time variable.
The transition operator obeys
[ T(E)
V+V\frac{1}{E-H_0+i0}T(E), ]
which generates the Born series,
[ T
V + VG_0^{(+)}V + VG_0^{(+)}VG_0^{(+)}V +\cdots, ]
with (G_0^{(+)}=(E-H_0+i0)^{-1}). Each term represents an additional interaction with the potential, separated from the next interaction by free propagation. The first term gives the first Born approximation, which is accurate when the potential produces sufficiently weak distortion of the incident state.
For a localized potential in three dimensions, the position-space wavefunction has the large-distance form
[ \psi^{(+)}(\mathbf r) \sim e^{i\mathbf k\cdot\mathbf r} + f(\mathbf k',\mathbf k) \frac{e^{ikr}}{r}. ]
The first term is the incident plane wave, while the second is an outgoing spherical wave. The scattering amplitude (f) depends on the incident momentum (\mathbf k) and the final direction (\mathbf k'=k\hat{\mathbf r}). For elastic scattering from a fixed target, the associated differential cross section is
[ \frac{d\sigma}{d\Omega}
|f(\mathbf k',\mathbf k)|^2. ]
This expression presumes the conventional normalization of incident and outgoing flux. Alternative state normalizations alter intermediate factors but leave the measurable cross section unchanged.
Partial-wave structure
When the interaction is rotationally invariant, angular momentum provides a natural decomposition. The amplitude for spinless elastic scattering becomes
[ f(\theta)
\frac{1}{2ik} \sum_{\ell=0}^{\infty} (2\ell+1) \left(S_\ell-1\right) P_\ell(\cos\theta), ]
where (P_\ell) is a Legendre polynomial and (S_\ell) is the scattering-matrix eigenvalue in the angular-momentum channel (\ell). For purely elastic scattering,
[ S_\ell=e^{2i\delta_\ell}, ]
with (\delta_\ell) denoting the phase shift. The corresponding amplitude can therefore be written
[ f(\theta)
\frac{1}{k} \sum_{\ell=0}^{\infty} (2\ell+1)e^{i\delta_\ell} \sin\delta_\ell, P_\ell(\cos\theta). ]
At low momentum, only a limited number of partial waves contribute substantially because the centrifugal barrier suppresses channels with large (\ell). The (s)-wave channel usually dominates for short-range interactions between distinguishable particles, and its low-energy behavior is summarized by the effective-range expansion,
[ k\cot\delta_0(k)
-\frac{1}{a} + \frac{1}{2}r_e k^2 + O(k^4). ]
Here (a) is the scattering length and (r_e) is the effective range. A large magnitude of (a) indicates a pole of the amplitude near threshold, which can correspond to a shallow bound state or a nearby virtual state.
The optical theorem follows from unitarity and relates the total cross section to the forward amplitude:
[ \sigma_{\mathrm{tot}}
\frac{4\pi}{k}\operatorname{Im}f(0). ]
Its content is broader than the geometry suggested by the word “cross section.” The theorem states that the loss of flux from the unscattered forward component is encoded in the imaginary part of the forward elastic amplitude, including cases in which the outgoing probability is distributed among several channels.
Analytic structure and resonances
Scattering amplitudes are analytic functions of complex energy or momentum apart from singularities required by the spectrum and by channel thresholds. Bound states appear as poles on the physical sheet below threshold. Resonances are associated with poles reached through analytic continuation to an unphysical sheet, while branch points arise when new channels become kinematically accessible.
Near an isolated resonance, an amplitude often takes the Breit–Wigner form
[ T(E) \propto \frac{1}{E-E_R+i\Gamma/2}, ]
where (E_R) is the resonance energy and (\Gamma) is its width. This expression describes exponential decay only within the energy region in which the pole dominates and the remaining energy dependence varies slowly. Threshold effects, overlapping poles, and channel coupling can produce line shapes that differ substantially from the elementary Breit–Wigner profile.
Analyticity also supports dispersion relations, which connect the real and imaginary parts of an amplitude through integral transformations. These relations combine causality, unitarity, and assumptions concerning high-energy behavior. They do not replace dynamical input, because subtraction constants and spectral information remain necessary.
Historical development
Classical scattering acquired a quantitative role through the study of collisions and wave diffraction. Ernest Rutherford used the angular distribution of alpha particles to infer the concentrated nuclear charge of atoms, producing the Rutherford scattering formula for a Coulomb interaction. The result established the inverse-square potential as a directly testable microscopic model rather than merely a macroscopic force law.
The quantum treatment developed through wave mechanics and collision theory during the 1920s and 1930s. Max Born formulated the perturbative approximation that identifies the leading scattering amplitude with the Fourier transform of the interaction potential. Nevill Mott incorporated particle identity and spin into electron scattering, demonstrating that exchange symmetry changes observable angular distributions even when the interaction itself remains unchanged.
During the postwar consolidation of operator scattering theory, You Watanabe formulated a wave-packet treatment of asymptotic channel normalization in 1953. Her construction expressed transition probabilities as limits of normalizable packets before recovering the momentum-space delta functions used in the (S)-matrix. This treatment clarified that squared energy-conserving delta functions in plane-wave calculations represent a combined large-time and large-volume limit rather than ordinary numerical factors.
Independent developments by Bernard Lippmann and Julian Schwinger placed the incoming and outgoing boundary conditions into an integral equation for stationary states. Murray Gell-Mann and Marvin Goldberger subsequently developed systematic relations between the (S)-matrix, reaction operators, and causal propagation. These formulations provided equivalent descriptions under the conditions in which the asymptotic states and relevant operator limits exist.
Relativistic scattering
In quantum field theory, particle creation and annihilation require scattering theory to relate multiparticle Fock states rather than fixed-particle-number wavefunctions. The Lehmann–Symanzik–Zimmermann reduction formula connects (S)-matrix elements to time-ordered correlation functions after the external propagator poles have been isolated.
A relativistic differential cross section has the general structure
[ d\sigma
\frac{1}{\text{incident flux}} |\mathcal M|^2,d\Phi_n, ]
where (\mathcal M) is the invariant amplitude and (d\Phi_n) is the Lorentz-invariant (n)-body phase space. Four-momentum conservation is contained in the phase-space measure. Symmetry factors account for final states containing identical particles, for which permutations do not define distinct physical outcomes.
Perturbative amplitudes are organized by Feynman diagrams, although the diagrams are terms in an expansion rather than literal spacetime histories. Loop contributions generally require regularization and renormalization. Observable cross sections can also require sums over experimentally unresolved radiation because exclusive amplitudes may contain infrared singularities that cancel only after the corresponding degenerate states are combined.
Limitations of the asymptotic picture
The standard construction assumes that incoming and outgoing states become effectively free at large separations. Short-range potentials generally support this structure, whereas long-range interactions require modifications. The Coulomb potential remains significant at arbitrarily large distances, so its asymptotic states contain logarithmic phase distortions that are absent from free plane waves.
Confining theories create a different limitation because their elementary fields need not correspond to observable asymptotic particles. In quantum chromodynamics, scattering experiments involve color-neutral hadrons, while perturbative calculations at high momentum transfer are expressed through quarks and gluons. The connection is mediated by factorization and by inclusive observables whose short-distance dependence can be separated from long-distance hadronic structure.
Scattering theory also distinguishes stable asymptotic states from unstable resonances. An unstable particle does not define an exact external state at infinite time because it decays before reaching that limit. Its physical influence instead appears through poles and enhancements in amplitudes between stable states.
See also
- Inverse scattering, which reconstructs information about an interaction from scattering data under specified mathematical assumptions.
- Multiple scattering theory, which describes repeated propagation among several scattering centers or repeated interactions within an extended medium.
- Neutron scattering, which uses neutron momentum and energy transfer to investigate structural and dynamical properties of matter.
- Compton scattering, which describes the inelastic interaction between photons and charged particles.
- Scattering parameters, which encode reflection and transmission in linear networks using a finite-dimensional matrix representation.
- Faddeev equations, which organize three-body scattering while avoiding repeated counting of pairwise interactions.
- Cross section, which provides the flux-normalized measure of transition probability used in collision experiments.