 |
The micrograph shows a 100 mm Si
wafer after preferential etching. The wafer was cut from a large as-grown
crystal and only polished before etching. |
|
 |
The crystal was grown with the
float-zone
technique and represented the state of the art in about 1972. The
typical spiral pattern of the small etch pits lead to the name "Swirl
defects". These defects were extremely detrimental to the functioning of
integrated circuits and power devices made from the wafer. It was thus of prime
importance to learn about their nature so that they could be avoided. |
|
|
|
|
|
|
|
The picture was taken under "dark field"
conditions. The wafer is illuminated at an angle; only light that is scattered
at defects reaches the lens of the camera. Perfect areas are totally black. The
defects must be due to agglomerates of the point defects (including perhaps the
major impurities O and C) that were present a high temperatures -
presumably in thermal equilibrium. |
|
The etch pattern at high magnifications as seen
through an optical microscope reveals two types of defects (see also the
pictures in the
link) The first picture is at an intermediate magnification, the second one
at high magnification: |
|
|
|
|
|
|
|
Lots of small etch pits can be seen in a striated
pattern - the swirl pattern. The inner areas of the wafer may only contain
these "B-type" defects, whereas closer to the edge of the
wafer, some large hillocks - the A-type" swirl defects are
contained within the B-defects. Hillocks and pits give different signs
of the black-white contrast (the vector from the black part of the contrast to
the white part); this serves to distinguish between the two possibilities. |
|
The a-type swirl defects are dislocations loops
and dislocation loop clusters of interstitial type -
the loops shown
before. This result was the first direct observation that showed that self
interstitials play a role in Si. Etching techniques can not provide a result
like that. |
|
|
In fact, it was never possible to establish the nature of the
B-type defects. They might be "fore-runners of the A-type
defects - i.e. some kind of interstitial agglomerate - or small vacancy
agglomerates; possibly small voids; but nobody knows for sure. |
|
|
Since present day crystals are much larger and grown with
different techniques, swirl defects are now longer seen. But other types of
defects (called C- and D-defects) are present now and always
first detected by optimized preferential etching solutions. D-defects
meanwhile have been identified as small voids, i.e. vacancy agglomerates |
|
|
|
© H. Föll
To index
Float Zone Crystal Growth