552 



THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1956 



Resistivities then measure only the dissolved lithium although the true 

 solubility at the temperature of saturation includes the precipitated 

 material. 



However, we have seen that germanium suitably doped with gallium 

 will not lose lithium by precipitation. Therefore the experiment might 

 be performed in doped germanium. The only difficulty with this sugges- 

 tion lies in the fact that doping chayiges the solubility. This objection can 

 be overcome through use of (3.4). In terms of that equation D'^ would 

 be measured in the presence of gallium whereas Do"^, the solubility in 

 undoped germanium, is required. But according to (3.4) if Z) , n, , and 

 A~ (gallium concentration) are known Do"*" can be computed. In fact 



solving (3.4) for Do yields 



^+ 



D^(D-^ - A-) 



+ 



Do"- = 



/ 



D^iD"- - A-) 



+ (Dyn,' 



V' 



rii + 



D^{D^ - A~) 



+ 



/ 



Z)^(D^ - A-) 



-\ 2 



(4.1) i 



+\2 2 



+ aryn^ 



The plan is therefore self-evident. Samples of germanium of known ! 

 suitable gallium contents A~ are to be saturated with lithium at various \ 

 temperatures. If a judicious choice of gallium content is made the lith- 

 ium will not precipitate when the specimen is cooled. Therefore the value * 

 of D^ characteristic of the saturation temperature can be determined ' 

 through resistivity measurements performed at room temperature. 

 Taking nj from Fig. 2 it then becomes possible to calculate Do using 

 (4.1). I 



The crystal specimens employed were cut in the form of small rec-l 

 tangular wafers of dimensions, approximately 1 cm X 0.4 cm X 0.1 cm. " 

 On the surfaces of these, small filings of lithium were distributed densely 

 enough so that their average separation was less than the half thickness 

 of the specimen's smallest dimension. The filings Avere alloyed to the 

 germanium specimen by heating in dry helium for 30 seconds at 530°C. ■ 

 Then the crystals w^ere permitted to saturate with lithium by diffusion 

 from the alloy at some chosen lower temperature. After the period of 

 saturation which ranged from one half hour to as long as 1G8 days, de- 



