Diffusion transistor#Mesa transistor

Under a Mesatransistor (English mesa transistor ) is defined as a set of bipolar transistors, are produced by a diffusion process. They therefore counts to the group of diffusion transistors. They differ from planar and epitaxial transistors with their manufacture and thus in shape, that is, in the topography, and in the manner how the differently doped regions are arranged. Mesa transistors were prepared in the beginning of germanium, which also has a higher charge carrier mobility than silicon. Had silicon in the 1950s, not today's dominance as the semiconductor material and came in at Mesatransistor lesser extent used. Nowadays mesa transistors are rarely used.

History

Although diffusion techniques have been used previously in the manufacture of transistors, only led the development of the oxide masking process by Carl frog and Lincoln Derick on the development of Mesatransistors, in 1954 by Charles A. Lee, GC Dacey and PW Foy, all employees at Bell Telephone Laboratories (BTL ) was invented. His name got the Mesatransistor by its appearance that resembles the mesas in Mesa (Arizona ) (Spanish mesa = table, see Mesa ( Semiconductor Technology ) ). The name itself should already have been named in 1954 by James M. Early, who led at the time the transistor development group at BTL. The Mesatransistor represents a significant advance over previous types of transistors, mainly because of the precision with which the depth of the exclusion zone could be controlled by diffusion. This has made it possible to reduce the width of the base region made ​​by diffusion to 5 microns and less, to about one -tenth of what was previously possible. This resulted in improved performance at high frequencies and an increase of the cutoff frequency of about 100 MHz.

Production

Mesa transistors were fabricated both npn and pnp transistors. As the semiconductor material used in the 1950s still germanium and silicon. In the following, the main steps for the manufacture of an npn transistor will be described, they are different in comparison to the manufacturing of a pnp transistor substantially only by the base materials and dopants employed. It should be noted that this is only a possibility of the production process from the early years in the description, which can be quite different in one or another part of step used later processes. For example, the selective masking material of thermal silicon dioxide was used for the diffusion of dopants.

The active regions of the npn Mesatransistors are typically formed on the surface of a p- doped single-crystal wafer. For that, a p-doped layer at the surface of the entire wafer has been generated by diffusing, for example, by the diffusion of gaseous arsenic in a diffusion furnace. The n-type substrate is the collector and subsequently the n- doped region, the base of the bipolar transistor dar. For the production of the emitter region diffusion step is further necessary. For this purpose a material is applied in a masked region on the p-doped layer, which leads to a diffusion step, a p-doped region, for example, aluminum. For now following diffusion, this arrangement is strongly heated. This not changed too much, the doping profile of the previous n-type doping during the process ( progress ), it is important that p- impurity atoms diffuse faster than the n- impurity atoms. It creates the typical layer sequence for a pnp pnp bipolar transistor.

For improving the electrical properties ( inter alia, the reduction of junction capacitances, improving the high -frequency power ) the top layers of the transistor are now etched down to the contact point of the wet chemical dopant materials with the semiconductor crystal. The etching of the sides gives the look of a transistor, the Table Mountain.

Then the individual portions of the layer sequence to be contacted, wherein the crystal contact to the collector and the other two are the base and emitter contacts.

Properties

The operating range of Mesatransistors is limited to a comparatively thin region at the surface. To maintain a high breakdown voltage between the base and collector of a relatively high collector resistance was necessary. There was attached at Mesatransistor the collector terminal is usually on the opposite side of the (relative to the active region ) thick substrate, was also high, the value of the collector bulk resistance. This results in a reduction of the power capacity of the component result. Attempts to reduce the substrate thickness, leading to stability problems of the semiconductor substrate and in a lower production yield. Solved the problems were only with epitaxial coating processes and the planar transistor.

The exposed portions of the collector- base junction on the surface to deteriorate the performance of the transistor. Caused by a high surface leakage current and low breakdown voltage. The unprotected against influences from the environment barrier layer also results in unstable transistor characteristics, causing a drift of the current blocking layer, the increase of leakage current and a reduction in the current gain. The reliability problems were quickly overcome for silicon transistors, as 1959, the group led by Martin M. Atalla ( also at BTL) showed that a thermally grown silicon dioxide layer passivates the surface. The layer could also be produced during the diffusion process of the base layer and emitter layer. The first transistor, who used this technique, however, was a oxidpassivierter bipolar planar transistor ( Fairchild Semiconductor 2N696, originally a Mesatransistor ) by Jean Hoerni.

The advent of the planar transistor has the Mesatransistor not completely eliminated. The Mesatransistor was still used when high breakdown voltages are required in the reverse direction of several thousand volts, such as power transistors.

In the design of the base and the emitter of two geometries are distinguished: a linear and a circular array. Wherein the circular array (English circular geometry ), the emitter electrode in the center of the annular base electrode. In the linear arrangement ( engl. linear / stripe geometry ), the two strip-shaped electrodes are parallel to each other (german 2 -stripe mesa ). The latter are typically employed for high frequency applications. In addition, they are simpler to design and requires less space.