Prism or grating couplers are known to exhibit m-lines in the far-fiel
d and zeros in the near-field. Despite the fact that a lot of papers h
ave been devoted to these couplers, there is some confusion concerning
the possible link between the m-line and zero: is the former a conseq
uence of the latter? One of the objects of this paper is to clarify th
is point. It is the opportunity to consider, in addition to the near-f
ield and far-field patterns, the question of the intermediate region c
onstituted by the Fresnel zone and to give a criterium of the far-fiel
d pattern.