SURFACE PROPERTIES OF GERMANIUM 



19 



n in p-type material and p in n-type. That this prediction is borne out 

 is shown by how well the experimental points fit the straight line curve 

 2 in Fig. 12 where (Ac.p.)i, is plotted versus this quantity. The scale 

 for 8p is shown along the bottom of Fig. 12 and that for tn(l -{- 8p/a) 

 along the top. Similar results were obtained for unit A. 



In Table I the parameters, specific resistance p in ohm cm, life time r 

 in microseconds and the surface recombination velocity Vs in cm/sec 

 are given for each unit used. Also given are some pertinent quantities 

 derived therefrom, namely the equilibrium densities of electrons and 

 holes, n and p in number per cm and the increase in density at the 

 surface 8p in number per cm^ when the rate of pair production due to 

 the illumination was 6.0 X 10^^ per cm^ sec. 



THEORY 



The constancy of Vs throughout the range of surface dipole investi- 

 gated puts rather stringent requirements on any theoretical model to 

 be constructed. W. Shockley and W. T. Read^ have investigated the 

 theory of recombination via traps. It is evident from their work that if 

 one assumes a trap density peaked near a single energy and very small 

 elsewhere, then Vs will be constant over a range of surface dipole values 

 provided that the peak energy is either near the conduction band or 

 near the filled band. The experimental results make it appear very 

 unlikely that the trapping mechanisms on the n and p-type surfaces 

 are essentially different. It is assumed that both types of traps are 

 present on the surface and that the traps are approximately the same 

 for both n- and p-type samples. Further, it is assumed that the traps 

 Na of energy Ea near the conduction band are donor type, i.e., neutral 

 when filled and positively charged when empty. Likewise the traps Nb 

 of energy Eb near the filled band are assumed to be acceptor-type traps, 

 i.e., neutral when empty and negatively charged when filled. The 

 absolute charge on the traps is not important, however, because we are 



Table I 



