SINGLE CRYSTAL BY ZONE LEVELING 651 



gradient is steep, the temperature of the liquid is above its liquidus 

 (freezing point curve) throughout the Hqiiid, and no stable nuclei can 

 form. However, increasing the growth rate decreases the temperature 

 gradient, while it depresses the liquidus. If the temperature gradient 

 is reduced to that indicated for fast grow^th, a region of constitutional 

 supercooling will exist in front of the solidifying interface where nuclei 

 can form and grow. The freezing of such a crystallite onto the growing 

 crystal marks the end of single crystal growth. 



A foreign body may also initiate polycrystalline growth. A natural site 

 for nucleation by foreign bodies is the wall of the boat, close to the growth 

 interface. Here the liquid germanium is in contact with foreign matter 

 at temperatures approaching its freezing point. It was found by D. Dorsi 

 that germanium single crystals could be grown satisfactorily in a smoked 

 quartz boat, at growth rates up to 2 mils per second. However, uniform- 

 ity considerations mentioned previously make it desirable to zone level 

 at much slower rates. 



It is believed that scattered dislocations may be produced in a single 

 crystal germanium lattice by three chief mechanisms. They may be prop- 

 agated from a seed into the new lattice as it grows; they may result 

 from various possible growth faults; but probably the most important 

 mechanism in this work is plastic deformation of the solid crystal. The 

 lirst cause may be minimized by selecting the most nearly perfect seeds 

 available, the second by using slow growth rates, and the third by mini- 

 mizing stresses in the crystal. 



The first hint that plastic deformation in the crystal might be an im- 

 portant source of dislocations came from the study of crystals pulled 

 from the melt by the Teal-Little technique. Frequently when sections of 

 crystals grown in the [111] direction were etched in CPi the pits were 

 arrayed in a star pattern, Fig. 8(a), in which the pits appeared on lines — 

 not randomly distributed. This coherent pattern suggested strongly that 

 the lines were caused by dislocations in slip planes which had been ac- 

 tive in plastic deformation of the crystal. The slip system of germanium 

 has been determined to be the <110> directions on {111} planes.^^ 

 If the periphery of the crystal is assumed to be in tension, it is possible 

 to calculate the relative shear stress pattern in each slip system of the 3 

 {111 { planes which intersect the (111) section plane. The results of these 

 calculations are summarized in Fig. 8(b) which shows a polar plot of 

 the largest resolved shear stresses for these planes and also their traces 

 in the section plane. The agreement with the observed star pattern is 

 striking. 



15 Treuting, R. G. Journal of Metals, 7, p. 1027, Sept., 1955. 



