In the impulse approximation (IA), which is used to interpret deep inelastic neutron scattering measurements, the predicted scattering is identical to that which would be obtained from a gas of noninteracting atoms, with the momentum distribution of atoms in the target system. The validity of the IA rest on the approximation that the struck particle recoils freely after the collision with the neutron. Departures from the IA are usually attributed to final-state effects (FSE), which are caused by the breakdown of this approximation. A second implicit assumption of the IA, which has received little attention in the literature, is that the atoms in the target system have a distribution of energies in the initial state, i.e., before the collision. This is not true in a quantum system at zero temperature, where, although there is a distribution of atomic momenta, the initial state has a unique energy. It is shown that ‘‘initial-state effects’’ (ISE), which are caused by the breakdown of the latter assumption at low temperatures, largely account for observed asymmetries and peak shifts from the IA prediction for S(q?,ω). It is shown that if FSE’s are negligible, ISE’s are negligible when q≫$p_i$, where q is the momentum transfer and $p_i$ is the root-mean-square atomic momentum. Finally, it is shown that FSE’s are negligible when q≫$p_i$, in systems other than quantum fluids and that, since in this regime ISE’s are also negligible, the IA is reached for such systems when q≫$p_i$. .AE

• Received 26 June 1989

DOI:https://doi.org/10.1103/PhysRevB.41.41