Higher-order breakdown in epitaxial P⁺N/N⁺ junctions

In the secondary breakdown characteristics of PN junctions operated in reverse bias, there is always one discontinuity when the reverse current in the avalanche breakdown region (first breakdown) reaches a critical value, I._SB1 at which the voltage drops suddenly to a lower value, V_SB1, In a few cases, more than one discontinuity was seen in devices operated at extreme current and voltage levels in P₊N/N₊ epitaxial junctions, These phenomena rarely occur compared to the occurrence of the second breakdown. They are found even in high quality junctions, but their probability is greater as the quality is reduced.There is some similarity between the two breakdown characteristics, but the difference between them is that in the second breakdown only one critical current and voltage drop occurs while in the other there is more than one. The experiment show that there is a difference between the internal processes in the two breakdowns, although a connection exists between them. Comparisons ore made between internal and external effects in both.breakdown types. The phrase suggested here for the second kind of breakdown is ‘ higher-order breakdown ’, although other phrases have been used (Schafft 1967).The experiments were performed in junctions described in the companion paper (Aharoni 1974). In these measurements the current was increased in the reverse direction until one or more voltage drops took place.‡ Simultaneously, the light intensity emitted from the microplasmas at the junction periphery (caused by avalanche multiplication) was measured (Aharoni and Bar-Lev 1971). To ensure good microscopic observation and measurements of the light intensity, contact was made with probes, instead of aluminium. Information was thus obtained concerning changes in current distribution in all breakdown regions and helped to explain differences between the two kinds of breakdown. The characteristics were also displayed by applying pulses of rectified sine waves to obtain reproduceable data from the same junctions. The microscope observations showed different phenomena in the two cases. In the second breakdown a single hot-spot was created in a predictable region, near the centre in circular devices, called here the ‘ natural ’ area. This name is used because thermal and electrical considerations show that the current concentration in this area increases as the external current increases and so a hot-spot can be expected there. This fact, along with the measurements, explains the differences and similarities between the two breakdown types. The conclusion is that only the last voltage level in the series of higher-order breakdowns can be considered to be the second breakdown, as has also been concluded by others for similar phenomena in junction transistors (Schafft 1967).It should be noted that since the higher-order breakdowns did not occur in all the junctions tested, and since most of the breakdowns were irreversible, it was necessary to perform many experiments on a large quantity of devices. Thus, the phenomena, divided into internal and external effects, are average behaviour and some deviations from this were found in single devices.