An investigation of the alloying technique for the fabrication of germaniump–n–ptransistors

Abstract

The base width of a transistor is an important factor affecting the cut-off frequency, current gain and maximum collector voltage. The paper is concerned with the manner in which wafer thickness, size of indium dots, alloying temperature, flatness of junction, furnace atmosphere and junction area affect the limits that can be achieved in the alloying process.(1) INTRODUCTION Fn order to examine the problems involved in the making of low-power (up to lOOmW) germanium p-n-p alloy-junction transistors, a number of experimental transistors have been made. Two of the most important of the electrical parameters of the transistor are the current gain, hf e , and the cut-off frequency, fa., and for these to have large values the distance W between the emitter and collector junctions must be small. Theoretically with small emitter current density and a base of uniform resistivity, a cut-off frequency greater than 15Mc/s requires a base width of less than IOJU. To achieve an effective spacing of this value without risk of local punch-through it is essential that the emitter and collector junctions are flat and parallel. A further requirement is that the /?-type germanium redeposited on cooling after alloying should be adequately thick and of suitable resistivity and lifetime. Consequently, the main part of the investigation has been to establish what is necessary in the alloying procedure to produce flat junctions and uniform redeposition.(2) MATERIALS USED The resistivity, p, and the lifetime, T, of the n-type germanium are two of the factors that determine the electrical characteristics of the transistor. The value of the collector-to-base breakdown voltage due to avalanche effect, V a , is approximately proportional to resistivity, 1 but collector-to-emitter punch-through voltage, V p , is proportional to W 2 lp. To ensure that the maximum collector voltage is as large as possible for the base width required, the resistivity should be chosen so that V a and V p are approximately equal. In addition, for small values of collectorto-base leakage current, I cb0 , and base series resistance, r bs , a small resistivity is required. The minority carrier lifetime must not be too small if an acceptable current gain is to be achieved. 2

Dilatancy, jamming, and the physics of granulation

“…Some of these phenomena are reported in [8] for a low-density gelling colloid in which strong attractive interactions are surely present, but below we report somewhat similar results for well controlled hard-sphere colloids at concentrations close to their glass transition. Similar observations have been made in suspensions of zeolite in amphiphilic solvents [11]. It is too soon to say whether or not these features are ubiquitous in all granulating suspensions.…”

“…These preliminary results are slightly ambiguous since we cannot rule out a slight increase in the colloid density during the collection of the initial granule by a self-filtration effect as reported in [9]. Nonetheless, they are consistent with reports of the melting of a granule by contact with a fluid droplet, even when both are drawn from a fluid of the same colloid density [11]. We hope to confirm or disprove this phenomenon in hard-sphere suspensions (rather than zeolite [11] or flocculated colloids [8]) in future work [10].…”

“…In fact, the granules created by the granulation process just described have fascinating properties whose physical exploration is frustratingly incomplete in the literature. We summarize here what can be gleaned from side remarks contained in papers primarily addressing other issues [8,9], and by private communication [11,12]. Some of these phenomena are reported in [8] for a low-density gelling colloid in which strong attractive interactions are surely present, but below we report somewhat similar results for well controlled hard-sphere colloids at concentrations close to their glass transition.…”

“…A similar and equally striking experiment involves placing such a melted droplet in contact with a granule that is still frozen. In this case, the frozen granule is rapidly coalesced and the result is a single, melted droplet of larger size [11]. Finally, a flowable spherical droplet can sometimes be converted back into a misshapen granule with a matt surface, simply by prodding it firmly with a spatula.…”

Jamming transitions in a schematic model of suspension rheology

“…This entails large capillary stresses which might maintain the jam; lumps of paste removed from a rheometer would then remain solid. Vibration could destroy the jammed structure and restore both the fluidity and a wet appearance, as is observed [7][8][9]. Rather than model all this directly, we focus here on the simpler case of stress-induced jamming within an idealised rheometer, with no free surfaces.…”