In
addition to measurements of single
structure dimensions for process control of semidonductor
devices it is important as well to measure
defect
densitiy on the whole
wafer. By knowing defect density and defect
distribution on
the wafer there is the possibility to set up compensation
strategies for increasing the yield. Most metrology
tools are not
suitable for such measuring tasks because
they are
far too slow for achieving a representative sample
of the whole
wafer. Defectoscopy in contrast to
e.g. Scatterometry
is able to
analyse a multitude of structures simultaneously.
Starting point for the method isba microscopic image
which of cause
yields no resolved structure data if size is below
wavelength. However
the benefit of defectoscopy is that isolated defects still
result in bright or dark pixels in the CCD-image bif
certain illumination conditions are fullfilled. The
reason for this is, that
isolated defects of an otherwise periodic grating
cause redistribution of electrical carges in these
elementar cells.In order to distinguish between different types of
defects,
illumination conditions can be varied.bThe response
on this variation can be used as a specific defect
signature. Defectoscopy images can be calculated with our simulation
tool "Microsim"
so at the same time a newsemiconductor processes are introduced,
procedures to detect new type
pf defects are available.

simulated
microscopic
image of a cross grating including a defect
[1]
|
S.
Rafler; T. Schuster; K. Frenner; W. Osten; U. Seifert: Improvements on
the simulation of microscopic images for the defect detection of
nanostructures, Proc. SPIE 6922 (2008) 692215 |
|