520 



THE BELL SYSTEM TECHNICAL JOURNAL, MAY 1954 



a surface layer on an 0.5-2.0 mil intrinsic layer, rather than a thin and 

 unsupported web. 



When operating biases are applied to a p-n-i-p transistor, holes in- 

 jected at the forward biased emitter diode diffuse across the n region of 

 the base then drift at high velocities through the field region to the 

 reverse biased collector p region just as in a PNP transistor. However, 

 in the p-n-i-p, the drift transit time through the collector field is com- 

 parable to the diffusion transit time through the base and contributes 

 to phase shift of the short-circuit current-transfer ratio, alpha. In ad- 

 dition, the emitter depletion layer capacitance, Cxe , which is unim- 

 portant in previous triodes, is relatively large in the p-n-i-p and degrades 

 performance at very high and microwave frequencies by providing a 

 low impedance shunt around the emitter junction. 



The details of structure and operation, design theory, a comparison 

 of p-n-p and p-n-i-p units and some experimental results are discussed 

 in the follo^^dng sections. The concluding summary reviews the theo- 

 retical and experimental work. 



STRUCTURE AND OPERATION 



Impurity Distribution 



In general, device characteristics depend on structure and on operat- 

 ing conditions. However, structure is more basic than operating con- 

 ditions. The spatial distribution of fixed charge centers (donors and ac- 



.— EMITTER LEAD 



— COLLECTOR LEAD 



Fig. 1 — Sectional view of a p-n-i-p transistor. 



