Database Reference
In-Depth Information
Front end (line)
Front end (line)
Back end (test & assembly)
Back end (test & assembly)
FEOL (Front end of line)
Film deposition
Lithography
Structuring
Doping
FEOL (Front end of line)
Film deposition
Lithography
Structuring
Doping
WT (wafer test)
EP (electrical parameters)
FT (functional test)
Assembly
Wafer thinning
Singularization
Packaging
Device test
Module assembly
Module test
WT (wafer test)
EP (electrical parameters)
FT (functional test)
Assembly
Wafer thinning
Singularization
Packaging
Device test
Module assembly
Module test
BEOL (Back end of line)
Film deposition
Planarization
Lithography
Structuring
BEOL (Back end of line)
Film deposition
Planarization
Lithography
Structuring
Fig. 2.6.
Illustration of the overall process flow and the various origins of data.
taken into account. Also, while nonnormal distributions can in principle be
accounted for properly, the procedure is cumbersome to implement and does
not immediately address the failure mechanisms that change the shape of the
distribution, for instance, to a multimodal distribution.
2.2.6 Process Experiment
Two lots of 25 wafers each were split identically into three groups at two
process steps (s. Fig. 2.7) to vary the process parameters of these steps and
in accordance the electrical parameters of certain devices. The intention of the
split was to vary the threshold voltages of both n- and p-type logic transistors
about the target voltage for each device. This is also called a performance
split, indispensable for dynamic performance characterization, as it results
in slow, nominal, and fast logic gates for the final product. As a side effect,
some parameters related to the threshold voltage (most notably saturation
current) and the so-called IO device coupled to the logic device will also
follow the split.
This particular experiment was chosen because its effects are well known
in advance and the analysis is reasonably tractable by conventional methods
(Fig. 2.8). Thus there is an established baseline to compare the results of our
newly developed data analysis methods against. We expect any successful
method to reconstruct the split information in the two individual lots and to
recognize that any residual differences between the two lots are not related
to the splits, as the splitgroups are identical. Further, both the intended
Search WWH ::
Custom Search