We present linear and circular polarization measurements of $3^{1}P\ensuremath{-}2^{1}S$ (501.6 nm) photons detected in delayed coincidence with electrons inelastically scattered from the
31P level of helium. Vector polarization and coherence parameters of the coincident radiation provide direct experimental evidence of the coherent nature of the excitation process. Atomic orientation is obtained directly from the circular polarization measurements. Analysis of the data also yields $\ensuremath{\lambda}$ and $\ensuremath{\chi}$ parameters for
31P excitations, which are compared with previous data and theory.
The coherence parameters associated with the electron impact excitation of the He 3 1 P state are studied for a Monte Carlo simulation of an experiment consisting of crossed electron and atom beams. The effects of changes in gas pressure and interaction region geometry are monitored by a fixed polarimeter analysing radiation resulting from the 3 1 P to 1 1 S (53.7 nm) transition. This work extends the simulations of Humphrey with the incorporation of a finite electron detector, a region of background gas and angular dispersion in the electron beam. Humphrey found that with an infinitesimal electron detector, the measured polarizations were not strongly dependent on interaction region geometry. In this work, however, it is found that the coherence parameters are very strongly dependent on interaction region geometry, pressure and the solid angle of the electron detector. Even very small solid angles can adversely affect the measured parameters. While angular correlations are not specifically studied, there are indications from simulations of electron scattering angle integrated polarizations that it is most important that any measurements involving a rotation of the photon detector should rotate about a symmetry axis of the gas source.
In this paper, we specifically examine the consequences of a finite scattering geometry on ''in-plane'' coherence experiments, i.e. polarization correlation or superelastic scattering experiments in which the incidence vector of the detected photon or the laser 'photon lies in the nominal scattering plane (/phi//nu/ = 0). The ''in-plane'' experimental configuration is adopted in order to give direct information about spin-orbit coupling effects. We show that a finite scattering geometry can give rise to severe distortion in the dependence of the coincidence or superelastic scattering signal intensity on the photon polarization and that this distorted behavior can be misinterpreted as arising from spin-orbit coupling effects. 11 refs., 3 figs.