| 摘要 |
756,339. Semi-conductor devices. WESTERN ELECTRIC CO., Inc. March 13, 1953 [March 14, 1952], No. 6978/53. Class 37. [Also in Group XL (c)] An electric signal translating device comprises a semi-conductor body having a region of high resistance intrinsic conductivity, a source for injecting carriers of one sign into the region, a drain for receiving these carriers and a grid means for controlling the flow of carriers by space-charge variation or by injecting carriers of the opposite sign into the region to control the injection of the first carriers. The material must have few carriers so that a space charge limited emission takes place between source and drain, and may consist of pure intrinsic material, or substantially intrinsic material with very few donor and acceptor centres which almost balance. In operation, the intrinsic region constitutes a space charge region and the device is an analogue of the thermionic valve. Sourcedrain current is controlled by variation of electric field, or space charge, or both, and the device is applicable to amplification, modulation or other translation of signals. In the intrinsic region there are substantially equal numbers of electrons and holes due to relatively profuse thermal generation of electron-hole pairs. In some embodiments, these carriers are swept out of the body by electric fields at a fast rate compared to their normal lifetime, so that the conductivity of the body is even less than that of ideally intrinsic material at the same temperature. Such pure material may be prepared by successive purifying processes as described in Specifications 632,942, 632,980, [Group XL (c)] until the donor-acceptor unbalance is small enough. The source means may consist of a heavily doped N + region in intimate contact with the body; in contrast to the thermionic cathode, such a source requires no heat and is capable of greater current densities. A zone of highly doped P + material corresponds to the negative control grid of the valve, and a further zone of N + material forms the drain, corresponding to the valve anode. Preferably the material comprises a single crystal, produced for example as described in Specification 727,678. The zones of N + and P + material may be produced by melting a donor or acceptor (e.g. indium or antimony) or alloy thereof, in contact with the body or as described in Specification 721,671. Germanium and silicon are preferred materials but lead sulphide, lead telluride and copper oxide are also specified. Donor or acceptor centres may be utilized to provide a fixed ionic space charge throughout the body, to get a high mu, or in particular regions such as the source grid region to reduce electron-hole generation in high fields, and for good transit time and low or negative output impedance. In germanium at room temperature, few donor and acceptor centres are neutralized by electrons and holes so that the " ionic space charge is substantially fixed regardless of the number of carriers, as distinct from the conditions in a gas discharge tube. In the case of silicon, at room temperature, a large number of the electrons are bound to donors so that by changing this number (e.g. by hot electrons energized by strong fields) the ." ionic " charge can be varied to simulate the effect of changing ion densities in gas discharge tubes; this effect is enhanced by having large numbers of donor and acceptor centres while keeping the difference (Nd - Na) small. Some applications employ material in which the lifetime of holes or electrons is specially controlled. This may be achieved by heat treatment or nuclear bombardment which modifies the number of " recombination centres " ; or by surface diffusion or addition of nickel (e.g. 1 mg. to 100 gms. of melted germanium). The use of silicon or silicon carbide reduces the tendency of hole accumulation owing to their wider energy gaps. The emission of carriers from the source. depends on the adjacent field and this is controlled by the potential of the control zone or grid. In addition the current to the drain can be varied by modifying the space charge produced by the injected carriers. This space charge is dependent upon the potential of the control grid and also the drain, and may also be modified by injecting carriers of opposite sign from the control zone or grid. Under conditions of space charge limited emission, a maximum occurs in the potential versus distance curve for an electron; current flow at this maximum can be calculated on the basis of Child's Law. Distribution of potential beyond the space charge layer is determined by boundary values and the space charge of the electrons themselves. The Specification describes how the current density, potential distribution, and transit time of carriers (electrons or holes) injected from the source may be calculated. The effect and nature of the space charge and the drift and diffusion portions of the current are considered. Design charts (Figs. 5-10) giving practical values and showing parametrical relationships, are provided. The tendency and effect of hole accumulation in front of an N + source, which produces a " stagnant region " adjoining a space charge region and methods of modifying this tendency (e.g. by making the source N + region and the adjoining portion of the space body of short hole-lifetime material or by illuminating the body with X-rays) are described, especially in connection with arrangements adapted to provide current gain. The fields and conditions necessary to provide " swept" " intrinsic material (i.e. where the normal carrier density is reduced by removal of the carriers) are discussed. The effect of Zener field is considered, reference being made to Specification 697,880. References are made to other publications. The present invention is distinguished from the " space charge " transistor of Specification 748,487. Embodiments.-Fig. 16 shows a signal translating device comprising a body 10 of high resistivity intrinsic silicon or germanium. P- type zones 13 and 14 constitutes the source, and drain respectively, and N zones 11 and 12 form the control grid. P zone 14 is biased negatively so that holes are injected from P zone 13 and flow to the drain zone 14. N zones 11, 12 may be biased slightly positive relative to emitter 13. The body material 10 is normally denuded of carriers, the holes and electrons produced by thermal generation being swept away to the zones by the electric field. Positive potentials applied to N zones 11 and 12 will reduce the electric field around the source so that fewer holes are injected and will also modify the space charge produced by the injected holes. The space charge may also be modified by applying a negative potential to N zones 11 and 12 so that electrons are injected from these zones into the intrinsic body. In this way input signals applied across resistance 17 modify the source-drain current to provide an output signal across resistance 16. Fig. 18 shows a modification in which electrons are injected from source N zone 110 through intrinsic layer 100 to drain N zone 120. The control electrode comprises zones 130 of P material arranged as a grid. Such a grid may be produced by boring the holes in the intrinsic material, introducing indium or an alloy of indium and the semi-conductor into the holes and diffusing the acceptor impurity by heat treatment. Alternatively, the P zone may consist of a band extending around the body, with or without the internal grid structure (Figs. 25, 26, not shown). Control is exercised as in Fig. 16. The space charge in the sourcegrid or the grid-drain region may be neutralized by having a few donors to make that part of the material weakly N type; this can provide favourable transit times. In a further modification (Fig. 20, not shown), the source and grid consist of alternate strips of N and P material in one face of the intrinsic. body, the drain N zone being on the opposite face. Fig. 21A shows a further modification in which a second control zone or grid 25 is included, in addition to control grid 130, source zone 110 and drain zone 120. The grid 25 may serve to reduce capacitance between grid 130 and drain 120 or may be utilized as a second control electrode for modulating or mixing signals. In a further modification (Fig. 22, not shown) a single source N zone is associated with two control zone-drain channels, an additional " barrier " zone of P- type which is connected to the source being arranged to divide the electron stream between the two channels. The device has two inputs and two outputs and may be used as a push-pull amplifier or mixer. A current gain device is described (Fig. 23, not shown) in which the signal is applied to a P zone which is normally biased slightly negative relative to adjacent portions of the body, so that when the P zone is made positive, holes flow towards the source tending to increase the electron current to the drain N zones. Fig. 24 shows a modification of the Fig. 23 arrangement, in which an auxiliary electrode 28 of N material is placed opposite the P zone 260 which injects the holes to increase the electron current from source N zone 210 to drain N zones 220. The signal may alternatively be applied to electrode 28 and the output may be taken from the source or drain zones. In a further modification, the N zone source, P zone control grid, and N zone drain are arranged as a series of coaxial cylinders (Fig. 27, not shown). To reduce heating, a lower current density emitter may be provided by 'adding a weakly P-type zone adjacent the N-type source zone (Fig. 28, not shown). Remote cut-off or volume-control characteristic may be obtained by non-linear spacing between the elements forming the control grid, or between the elements and the source (Figs. 29 and 30, not shown). Cooling may be effected by having a mass of copper fixed to the drain electrode (Fig. 15, not shown) |
