Improvements in semiconductor devices

Abstract

Abstract 945,249. Semi-conductor devices. GENERAL ELECTRIC CO. Sept. 6, 1960 [Sept. 8, 1959], No. 30695/60. Heading H1K. A multi-layer semi-conductor device comprises at least two over-lapping NPN and PNP transistor units, a first ohmic electrode being connected to one terminal zone and a second ohmic electrode connected across the PN junction of the other terminal zone and its adjacent zone. Fig. 1 shows an NPNP arrangement with electrode 13 ohmically connected to P-zone 4 and electrode 12 ohmically connected to both N-zone 6 and P-zone 5 so that it shortcircuits the PN junction J E1 . The central PN junction J c between P-zone 5 and N-zone 3 behaves as a collector while the other two PN junctions J E1 and J E2 operate as emitters. If electrode 13 is biased negatively, the reversed biased junction J E2 breaks down at voltage 20, as shown, by the third quadrant characteristic in Fig. 2. If electrode 13 is biased positively, junctions J E2 and J E1 (partially) are biased forward and junction J E in reverse. At low currents, short-circuited junction J EI is inoperative as an emitter, but as the bias increases junction J C breaks down at voltage V BO and the increased current results in forward biasing of the right hand edge of junction J E1 away from the short-circuited point 11, so that emission takes place and alpha increases. The device then switches over to its good conducting state at a low voltage 23, as shown in Fig. 2. The switching current Is may be greater or less than the holding current I H , and curves showing the variation of these characteristics with temperature and emitter and base electrode widths are given. In a modification (Fig. 6, not shown), electrode 13 short-circuits junction J E2 ; then when 13 is biased positively the characteristic is like that for Fig. 1, but when 13 is negative the arrangement is merely a forward biased PN junction. Fig. 8 shows a symmetrical NPNPN arrangement having two collector junctions J c1 and J c2 ; the characteristic consists of the first quadrant of Fig. 2 for both negative and positive bias to electrode 13. Fig. 13 shows an arrangement similar to Fig. 1 but with an additional electrode 19 ohmically connected to the upper P-zone. Potential applied to this electrode enables the values of Is and V 10 to be controlled. Alternatively, the additional electrode can be applied to the zone lying on the opposite side of the J C junction. Fig. 16 shows an arrangement resembling Fig. 1 in which the N-zone 6 is split into separate portions avoiding the need for large injecting currents from electrode 26 to effect control. In further modifications additional electrodes are provided to the various zones for control purposes. The arrangement may be used for switches, oscillators, controlled rectifiers or negative resistance devices. The devices may be constructed by oxidizing the surface of a 3 ohm cm. N-type silicon block and heating the block in a tube containing silicon, gallium and phosphorus. The gallium, but not the phosphorus, diffuses through the oxide layer to form a P-zone; in regions free of oxide the gallium diffuses deeper than the phosphorus resulting in the formations of a surface N and underlying P-region. Unwanted layers are removed by grinding and etching. Gold is evaporated on to the surface to form electrode 12 and aluminium to form electrode 13. Nickel or lead may alternatively be used. A stream of boron chloride or boron oxide may be used to provide P-type regions. In Fig. 13, aluminium electrode 36 is mounted on a tungsten plate 34 soldered to a mounting conductor 35. Similarly, conductor 39 is soldered to tungsten plate 38 joined to aluminium electrode 37. Electrode 19 consists of aluminium wire.

Inventors

No inventors recorded.

Bibliographic Details