HOLE ISJECnOS IX GERM AM CM 345 



(luctors. It is worth while at the outset to contrast some of the new aspects 

 of these experiments with the earlier experimental status of the bulk proj)- 

 erties of semiconductors. I'rior to the invention of the transistor, inferences 

 about the behaviors of holes and electrons were made from measurements 

 of conductivity and Hail efTecl. l-'or both of these effects, under essentially 

 steady state conditions, measurements were made of such cjuantities as 

 len<^ths. currents, voltages and magnetic fields. 'J'he measurement of time 

 was not involved, except indirectly in the calibration of the instruments. 

 .Nevertheless, on the basis of these data, definite mental pictures were 

 formed of the motions of holes and electrons describing in particular their 

 drift velocity in electric fields and the transverse thrust exerted upon them 

 by magnetic fields. The new experiments show that something actually does 

 drift in the semiconductor with the predicted drift velocity and does behave 

 as though it had a plus or minus charge, just as expected for holes and 

 electrons. In addition, experiments described elsewhere show that the effect 

 of sidewise thrust by a magnetic field actually is observed in terms of the 

 concentration of holes and electrons near one side of a filament of germanium. 

 We shall discuss here evidence that holes are actually introduced into 

 //-type germanium by the forward current of an emitter point and show how 

 the numbers and lifetimes of the holes can be inferred from the data. We 

 shall refer to this important process as "hole injection." Discussions of the 

 reasons why an emitter should emit holes are given for point contacts by 

 J. Hardeen and W. H. Brattain ' " and for p-n junctions elsewhere in this 

 journal. There are other possible ways in which semiconductor amplifiers 

 can be made without the use of hole injection into «-type material or electron 

 injection into />-type material.* In this paper, however, our remarks will be 

 restricted to semiconductors which have only one type of carrier present in 

 appreciable proportions under conditions of thermal equilibrium; for such 

 cases the theoretical considerations are simplified and are apparently in good 

 agreement with the e.xperiments. 



2. The Measurement of Density and Current of Injected Holes 



The experiment in its semiquantitative form is relatively simple and is 

 shown in Fig. 1.^" A rod of «-type germanium is subjected to a longitudinal 

 electric field E applied by a battery Bi. Collector and emitter point contacts 

 are made to the germanium with the aid of a micromanipulator. The col- 

 lector point is biased like a collector in a type-A transistor by the battery 

 B>, and the signal obtained across the load resistor R is applied to the input 

 of an oscilloscope. At time /i the switch in the emitter circuit is closed so 

 that a forward current, i)roduced by the battery B-^, tlows through the emitter 

 point. At /s the switch is opened. The voltage wave at the collector, as 



* For example see references 1 and 1 1 . 



