Cylindrical symmetry discrimination of magnetoelectric optical systematic effects in a pump-probe atomic parity violation experiment

A pump-probe atomic parity violation (APV) experiment performed in a longitudinal electric field , has the advantage of providing a signal which breaks mirror symmetry but preserves cylindrical symmetry of the set-up, i.e. this signal remains invariant when the pump and probe linear polarizations are simultaneously rotated about their common direction of propagation. The excited vapor acts on the probe beam as a linear dichroic amplifier, imprinting a very specific signature on the detected signal. Our differential polarimeter is oriented to yield a null result unless a chirality of some kind is acting on the excited atoms. Ideally, only the APV ( -odd) and the calibration ( -even) signals should participate in such a chiral atomic response, a situation highly favourable to sensitive detection of a tiny effect. In the present work, we give a thorough analysis of possible undesirable defects such as spurious transverse fields or misalignments, which may spoil the ideal configuration and generate a chiral response leading to possible systematics. We study a possible way to get rid of such defects by performing global rotations of the experiment by incremental angular steps , leaving both stray fields and misalignments unaltered. Our analysis shows that at least two defects are necessary for the -odd polarimeter output to be affected; a modulation in the global rotations reveals the transverse nature of the defects. The harmful systematic effects are those which subsist after we average over four configurations obtained by successive rotations of 45 . They require the presence of a stray transverse electric field. By doing auxiliary atomic measurements made in known, applied, magnetic fields which amplify the systematic effect, it is possible to measure the transverse E-field and to minimize it. Transverse magnetic fields must also be carefully compensated following a similar procedure. We discuss the feasibility of reducing the systematic uncertainty below the one percent level. We also propose statistical correlation tests as diagnoses of the aforementioned systematic effects.Received: 19 November 2003, Published online: 19 February 2004PACS: 32.80.Ys Weak-interaction effects in atoms - 32.60. + i Zeeman and Stark effects - 33.55.Fi Other magnetooptical and electrooptical effects - 42.25.Lc Birefringence