“…To overcome the difficulty of using a published data set for a prediction or a simulation of SEM images, the best strategy would be to perform in-situ measurements of the SEE yields of specimen's components. This experimental strategy may consist in implementing, into scanning microscopes, attachments derived from that used in dedicated SEE measurements [6,16] such as that described by Iyasu et al [9] for conductors or, with slight changes, that of El Gomati and Assa'd [24] for any kind, conductive or insulating, of specimen. This strategy would be applied to specimens composed of a few components such as those involved in the research and developments of microelectronic systems and it may be limited to acquisition of only one measured pair of values, [d max ; E1 max ], for each component, these experimental values being inserted next into calculated-reduced yield curves, in order to obtain a normal yield curve, d ¼ f(E1) [19].…”