INTERFERENCE MEASUREMENT OF THIN SURFACE LAYERS 1213 



of both readings compensates the error. To obtain, however, the maxi- 

 mum of accuracy, one can locate the junction at any point. (See Fig. 3, 

 Point A). IVIoving the sample in the y-direction scribes a line through 

 the point at which the junction was found. A movement in the x-direc- 

 tion with the needle in the acjuadag marks a point B on this line. The 

 distance from this point to the junction can be obtained from the read- 

 ings on the micrometer. Thus, the exact point at which the junction was 

 located can be reproduced under the microscope. 



DETECTION OF THE p-n JUNCTION 



1. Thermoelectric Probe 



The thermoelectric voltage^ occurring between a hot and a cold con- 

 tact to the sample, changes sign by crossing the junction with the hot 

 contact. The advantage of this probe is that it does not depend upon 

 the rectification properties of a p-n junction. The thermoelectric probe 

 is most suitable for germanium since lapping across a p-n junction nor- 

 mally produces a "short" between the two regions. However, it is likely 

 to give a p-reading on lightly doped n-material. It is therefore only 

 usable on heavily doped layers, where the nearly compensated zone is 

 very small. In the case of silicon, the junction normally maintains recti- 

 fying properties after lapping; thus, a photocurrent is present. This cur- 

 rent is superimposed upon the thermocurrent. Therefore, the thermo- 

 electric probe is only usable in the dark. The photoelectric method (see 

 below) is more convenient for these cases. 



The thermoelectric probe used, consisted of a commercial phonograph 

 needle, which had a good hemispherical point and was surrounded by a 

 piece of ceramic tubing carrying a heating coil. Between needle and sam- 



Fig. 3 — Schematic view of a scribed p-n junction. 



« V. A. Johnson and K. Lark-Horowitz, Phys. Rev., 69, p. 259, 1946. 



