4 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1953 



and occupancy of the traps compensate in their effect on surface re- 

 combination, so that Va is unchanged by ambient. This means that 

 changes in concentrations of electrons and holes in the interior with 

 illumination and thus the body contribution to (Ac.p.)l are independent 

 of ambient. Changes in (Ac.p.)l with ambient result soley from changes 

 in Vb and are in the directions we have described earlier. 



The concentrations of carriers, n and p, and their change with light 

 are obtained as follows. Drift mobilities, Hn and jip , and the equilibrium 

 product, np, are knowTi from earlier experiments of G. L. Pearson, J. R. 

 Haynes and W. Shockley.^ From these, and the resistivity of the sample, 

 which was measured, n and p can be determined. The light source is 

 calibrated in terms of hole-electron pairs created per cm per sec. 

 The recombination rate is determined from the body life time r and the 

 surface recombination velocity. From these latter three measurements, 

 one can calculate the steady state density of electrons and holes near 

 the surface when it is illuminated. 



Mention should be made here of the fact that oxygen has been found 

 to play a definite role on the Ge surface. The large extreme in dipole is 

 obtained when active oxygen (ozone) is introduced into the gaseous 

 ambient. Peroxide vapors have the same effect. The other extreme is 

 produced by vapors having an OH radical, water vapor, alcohol etc. A 

 number of vapors not falling in either of the above classes have little 

 or no effect on the surface dipole. Another result is that the difference 

 in work function or dipole between n- and p-type Ge is small. This is 

 to be expected from previous work. 



We shall first discuss the experimental technique and then give the 

 experimental results. The main conclusions of the theory will then be 

 outlined and compared with these results. 



EXPERIMENTAL METHOD 



The Ge surface to be measured and the reference electrode are mounted 

 under a bell jar. Oxygen or nitrogen as desired is allowed to flow through 

 the bell jar at a rate of approximately 2 liters per min. The volume of 

 the bell jar is 16 liters. The gas used flows over a drying column of silica-gel 

 and then calcium chloride. Means are provided for bubbling this gas through 

 any desired liquid before it enters the bell jar. A spark discharge can be 

 run in the gaa flow line. The reference electrode is placed parallel to the Ge 

 surface about 1 mm away. It is mounted on a vibrating reed which is 

 driven, electromagnetically, at its resonant frequency of about 90 cycles 

 per second and at an amplitude of the order of 0.1 mm. This varies the 

 capacity sinusoidally giving rise to an electrical signal when any potential, 



