It has been experimentally found that it will be difficult to grow 300 mm o
r larger diameter crystals with similar quality as for 200 mm or smaller di
ameter crystals. This phenomenon can be understood within the frame of the
Voronkov theory in which the value of the parameter V/G (V = pull rate, G =
temperature gradient at the growth interface) determines which type of def
ect forms in the growing crystal. Due to fundamental technological constrai
nts, the pull rate of silicon single crystals has to be reduced as the diam
eter increases. For crystal diameters beyond 300 mm, the reduction of pull
rate is so large, that V/G(r) (r = radial position) can probably no longer
be kept above the critical value C-orit = 1.34 x 10(-3) cm(2) K-1 min(-1) o
ver the entire crystal volume by the present growth technology. As a result
, the defect behavior of the silicon bulk changes. The aggregation of defec
ts is now dominated by excess Si interstitials instead of vacancies and, he
nce, L-pits (dislocation loops) are observed instead of microvoids. Unless
new methods for the suppression of L-pits can be developed, this will serio
usly challenge the use of polished wafers in very large diameter device man
ufacturing lines, as L-pits can severely damage the device performance. A p
romising solution to the defect problem appears to be p+p- epi wafers. (C)
1999 Elsevier Science B.V. All rights reserved.