The electronic world has transformed from one that was almost exclusively analog to one that's now nearly all digital. The lengthy analog interlude at the beginning of the twentieth century makes it easy to forget that electronics started out as digital. You could look all the way back to the telegraph and electromechanical components such as relays, which are definitely are digital. These are probably better defined as "electric," rather than "electronic," since electronics is more than passing a current through a wire. It involves an active element, such as a vacuum tube.
The first digital electronic element was the coherer, a radio detector that replaced the primitive spark gap. The coherer, which is essentially a tube filled with metal powder with electrodes at each end, was invented by the French physicist, Édouard Branly, around 1890. The tube has a reasonably high resistance, but when it's connected between ground and a radio antenna, the resistance will decrease upon reception of a radio pulse. With appropriate circuitry, a coherer could both detect and amplify a radio signal so it could be heard. Unfortunately, the radio operator needed to tap the device with a hammer every once in a while to regenerate its action. This lead to the invention of an automatic tapper that would hammer the tube after receipt of each radio pulse. The operating principle of the coherer is probably the dielectric breakdown of the oxide that coats the metal powder; but, as you can imagine, there's not enough research interest in the coherer to elucidate its operating principle.
The first vacuum tube, the De Forrest Audion, was also a digital device. The audion was a tube containing a filament and a plate, just like a vacuum tube diode, but it was filled with a very low pressure gas. There was a wire wrapped around it that was connected through a tuned circuit to an antenna. Radio waves would cause an increased electrical conductivity between the filament and the plate, so the device acted like a coherer radio detector. Like the coherer, its operation was non-linear, so it could be considered a digital circuit element. Audions were later replaced by triodes and other vacuum tubes that contained no gas and were linear amplifiers. At that point, around the second decade of the twentieth century, electronics became mostly analog. Analog reigned supreme through most of the century until the transistor made digital electronics practical in the form of RTL, DTL and
TTL logic circuits that presaged the modern computer.
Fig. 4 from Lee De Forrest's Audion Patent (Ref. 1).
The transistors in computers act as on-off switches. A proper assemblage of such switches can make logic circuits for such functions as NOT, OR, AND, and the like. Such transistors are just an elegant version of the electromechanical relays that powered some of the first demonstration logic devices created by George Stibitz at Bell Labs. Since it's difficult to make transistors that operate at very high temperatures, engineers at Case Western Reserve University (Cleveland, OH) decided to make a type of high temperature relay instead.[2-6] The relay is made from silicon carbide, and it doesn't use magnetic force actuation like conventional relays. It's a NEMS (nanoelectromechanical system) device that uses electrostatic force for actuation. Just like an insulated-gate field effect transistor, there's no current required on the input electrode to hold its state, and there's a switching action from a zero conductance to a highly conductive state.
To make the SiC relays useful in electronic circuitry, the Case team configured pairs of these as a complementary switching cell that operates much like the CMOS cell it would replace. The switching cell, shown in the figures, converts an input voltage to an inverted output voltage, and it's been operated up to 500oC. The actuation speed is somewhat slow, only 500 kHz, but these structures have been operated for more than 2 billion cycles at 500oC. Five hundred kilohertz might seem too slow to do something useful, but the original RadioShack computer, the TRS-80, had a clock speed of only 1.77 MHz, and it was suited for many tasks.
Schematic diagram of an SiC NEMS complementary switch cell.
Cross-sectional view of an SiC NEMS complementary switch cell.
The simple structure shown in the figure ensures complementary action. However, there will be either a high impedance transition state between the logic states, or a short circuit if both cantilevers contact the output at the same time. This can happen in a mechanical system if resonance occurs at a particular operating frequency.
References: