Alignment method and semiconductor device

Abstract

The alignment method allows a constant decision of an alignment point even with an indistinct outline of an alignment mark. An operator moves a chip (2) so that the whole or parts of an alignment mark (1) (including at least angles D0A0B0 and A0B0C0) is included within a lens view field (4), and then decides an alignment point (AP0). More specifically, the operator reads angles which are specified to obtain bisectors, from a defect inspection apparatus; obtains respective bisectors of the angles; and decides the intersection thereof to be the alignment point (AP0). Then, a stage drive required to superimpose the alignment point (AP0) on the center (O) is calculated on the basis of the shift amount between the position of the alignment point (AP0) and the center (O) of a target scope (5) displayed at the lens view field (4). The chip (2) is then moved by the stage drive.

Dependencies are inferred from the claim text.