474 BELL SYSTEM TECHNICAL JOURNAL 



input impedance will be low, the power gain formula a R22/Rn will yield 

 power gain even when a is less than unity. 



In certain ways the structure of Fig. 10 resembles a vacuum tube. In Fig. 

 11, we show the energy band diagram, with energies of holes plotted upwards 

 so as to be in accord with the convention for voltages, (a) shows the thermal 

 equilibrium distribution and (b) the distribution under operating condi- 

 tions. It is seen that the potential hill, which holes must climb in reaching 

 the collector, has been reduced by ip^ . The w-region represents in a sense the 

 grid region in a vacuum tube, in which the potential and hence plate current, 

 is varied by the charge on the grid wires. Here the potential in the w-region 

 is varied by the voltage applied between base and emitter. In both cases one 

 current is controlled by another. In the vacuum tube the current which 

 charges the grid wires controls the space current. Because the grid is negative 

 to the cathode, the electrons involved in the space current are kept away from 

 the grid while at the same time the electrons m the grid are kept out of the 

 space by the work function of the grid (provided that the grid does not 

 become overheated.) In Fig. 11, the electrons flowing into the base control 

 the hole current from emitter to collector. In this case the controlled and 

 controlling currents flow in the same space but in different directions because 

 of the opposite signs of their charges. 



As this discussion suggests, it may be advantageous to operate the p-n-p 

 transistor like a grounded cathode vacuum tube, with the emitter grounded 

 and the input applied to the base. 



The p-n-p transistor has the interesting feature of being calculable to a 

 high degree. One can consider such questions as the relative ratios of width 

 to length of the «-region and the effect of altering impurity contents and 

 scaling the structure to operate in different frequency ranges. However, 

 we shall not pursue these questions of possible applications further here. 



Acknowledgment 



The writer is indebted to a number of his colleagues for stimulating dis- 

 cussions and encouragement, in particular to H. R. Moore, G. L. Pearson and 

 M. Sparks, whose experhnental work, to be described in later publications, 

 suggested development of the theory along the lines presented above. He is 

 also indebted to J. Bardeen, P. Debye, G. Wannier and W. van Roosbroeck 

 for theoretical comments and suggestions and especially to the last and 

 to Mrs. G. V. Smith for valuable assistance in preparing the manuscript. 



" Physical Principles Involved in Transistor Action, J. Bardeen and W. H. Brattain, 

 Phys. Rev. 75, 1208 (1949). 



