Tuesday, September 29, 2020
Was this new or just deceptive?
An intriguingly vague innovation in a 1960 issue of Westinghouse Engineer, newly added to American Radio Library.
This was clearly one way of making ICs, but it doesn't sound the same as the path already started by Intel in 1958. The Intel method resulted in identifiable components formed as part of the same layered substrate by controlled etching.
This Westinghouse proposal sounds more like the whole substrate is participating, with loosely defined areas serving as complete functions, not individual R and C and transistor. Later iterations of ICs came closer to this idea, but still have mainly distinct and identifiable components.
This involves a new concept in the design and function of electronic systems, called "molecular electronics," to indicate its dependence on phenomena occurring within or between domains of molecules in the solid state. As one accomplishment of the joint program, a variety of molecular electronic "function blocks" are being produced, three of which are shown in Fig. 1; these solid-state elements achieve, entirely within themselves, electronic results previously gained only by assembling many varied items of electronic hardware. Because of this, these elements are not intended as "components," such as transistors and tubes, but rather as "subsystems." Examples of functions performed by function-blocks are such electronic operations as amplification, oscillation, and telemetering. Effects of this general type are used in molecular electronic blocks by creating - usually in single crystals - a number of distinct operative domains, which can be regarded as molecular "communities" having a common civic purpose, in that each domain will sustain a desired electronic occurrence. The domains border one another at boundaries called interfaces, which are like political frontiers in their ability to initiate phenomena different from those occurring inside the molecular domains. As a simple example, the element diagrammed in Fig. 3 is composed of two domains that meet physically at one interface. One of these domains is composed of a resistive material selected and shaped to present a resistance R1 to the passage of current; the other domain is also resistive, but is so planned that it has a resistance R2. At the interface, the interaction between domains causes a capacitive effect. Thus, this tiny element is a subsystem equivalent to a time-delay circuitWas this just a poorly worded description of the Intel process, or an intentionally deceptive description for competitive and patent purposes, or an entirely different concept? Hard to tell. The idea of custom-shaping a component to serve an entire function was already familiar before ICs, in the microwave realm where every wire and transistor and space between components is unavoidably acting as inductor and capacitor. The entire system needs to be designed physically and geometrically all at once, more like a sculpture than a schematic. The article doesn't start from this familiar concept.
Labels: Asked and unanswerable
¶ 5:08 AM
