612 THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1956 



zinc or gold. To some extent the methods of this paper can be appHed 

 to these. Furthermore, returning to Hthium, there are impurities both 

 mobile and immobile which introduce more than one energy state into 

 the forbidden gap. The phase relations of lithium in the presence of these 

 should be extremely interesting since the corresponding mass action 

 equations are more complicated. Analogues of dibasic*^ acids and bases 

 should exist. 



In the case of ion pairing doubly charged acceptors like zinc in ger- 

 manium^^ should be extremely interesting, since large amounts of pairing 

 should persist up to very high temperatures. In fact such studies repre- 

 sent excellent means of testing for the existence of doubly charged ions. 

 There is also the question of what happens to the two energy levels when 

 an acceptor like zinc pairs with a single lithium ion. Are both levels 

 driven from the forbidden gap or do they split under the perturbation? 



Then there is the problem of ion triplets — a possibility with impuri- 

 ties hke zinc ■ — which is unexplored both theoretically and experiment- 

 ally. Also, very strange diffusion effects must occur in the presence of 

 doubly charged ions, to say nothing of the effect which uncompensated 

 mobile holes might have on relaxation processes. 



The field of ion pairing in silicon is relatively unexplored. 



All of the phenomena discussed in this paper must occur in the group 

 III-V compounds, more or less complicated by additional effects. 



The question of the formation of the LiB~ complex in both germanium 

 and silicon needs further study. It should behave as an acceptor and its 

 electronic energy state might be revealed by suitable quenching 

 techniques. 



Non ionic reactions between group V donors and group III acceptors 

 very likely occur, i.e., a real III-V covalent bond may be formed be- 

 tween such atoms dissolved in germanium or silicon at high tempera- 

 tures. This possibility could be investigated by looking for changes in 

 carrier mobility or impurity energy levels upon extended heating — in 

 much the same way that ion pairing has been studied. If found, the 

 phenomenon may provide an excellent technique for measuring the dif- 

 fusitivities of all classes of impurities even at fairly low temperatures. 



Such compounds may possess strange energy levels and be responsible 

 for unexplained traps and recombination centers. 



The effect of stress on the extent of ion pairing may well be profound 

 since there will be a tendency for such stress to concentrate at imperfec- 

 tions. Stress studies on ion pairing may therefore be useful for further 

 investigating the strain about an isolated impurity. 



Ion pairing between lithium ions and acceptor centers located in 

 dislocations or vacancies should occur. In the first case the dislocation 



