This work deals with the computation of electron pair correction to small a
ngle Bhabha scattering, in order to contribute to the improvement of lumino
metry precision at LEP/SLC below 0.1% theoretical accuracy.
The exact QED four-fermion matrix element for e(+)e(-) --> e(+)e(-)e(+)e(-)
, including all diagrams and mass terms, is computed and different Feynman
graph topologies are studied to quantify the error of approximate calculati
ons present in the literature.
Several numerical results, obtained by a Monte Carlo program with full matr
ix element, initial-state radiation via collinear structure functions, and
realistic event selections, are shown and critically compared with the exis
ting ones.
The present calculation, together with recent progress in the sector of O(a
lpha(2)) purely photonic corrections, contributes to achieve a total theore
tical error in luminometry at the 0.05% level, close to the current experim
ental precision and important in view of the final analysis of the electrow
eak precision data. (C) 1999 Elsevier Science B.V.