Dual Implants for Etch Stops

Dual Implants for Etch Stops

       Wafer bonding is a promising method to create
silicon-on-insulator starting material.  Two oxidized Si wafers are
bonded together oxide to oxide and one of them mechanically and
chemically removed until a thin film of Si remains, separated from
the bottom wafer by the combined oxides.  For CMOS and other advanced
technologies, the Si layer should be very thin, from 30 nm to several
hundred nm.  This thinness is difficult to achieve with good
thickness control.  One method to accomplish this is to create a
heavily doped p-type region in the first wafer, followed by implant
annealing, if necessary, grow an epitaxial layer on the p+ surface,
oxidize the epi layer, then use this structure as the upper wafer in
the bonding process.  The this part can be mechanically removed, and
a preferential etch used to remove the remainder down to the p+
surface.  The p+ region is then removed to leave the epi-layer as the
remaining active Si region for device formation.

      The acceptor species used for p+ layer formation is B, which is
usually implanted at doses high enough to create greater than 7x1019
cm-3 for the selective etch to function.  During this implant, a high
number of defects are created in the Si which coalesce to form
predominantly dislocation loops in the implanted region upon
annealing.  These defects act as nucleation sites for dislocations
which propagate into the epitaxial layer during epi growth and are
therefore present to some extent in the finished device layer.  This
problem is caused by the failure of B implantation to create an
amorphous layer which can be annealed out with very few remaining
defects.

      The necessary features of > 7x1019 cm-3 doping with low defect
density can be accomplished by implanting with an amorphizing species
prior to the B implant.  The amorphou...