By the mid 1970s, the only vacuum tubes you could find in western electronics were in certain kinds of specialized equipment. Currently, vacuum tubes are pretty much a nonexistent technology, but that may change in the future. Some changes to the fabrication techniques used in integrated circuit design could bring vacuum electronics back.
NASA Ames Research Center has been working to develop vacuum-channel transistors. While the research is still in the early stages, their prototypes hold great promise. Vacuum-channel transistors have the potential to work 10 times as fast as ordinary silicon transistors and may have the ability to operate at terahertz frequencies. They are also much more tolerant of heat and radiation. To understand why these development may be possible, it will help to understand a little about the construction and functionality of vacuum tubes. While the vacuum tubes that amplified signals in radios and televisions during the first half of the 20th century seem to not resemble the metal-oxide semiconductor field-effect transistors (MOSFETs) that are used in modern electronics, they do have similarities. Both are three-terminal devices. The voltage applied to one terminal, the grid for the vacuum tube and the gate for the MOSFET, controls the amount of current flowing between the other two (from cathode to anode in a vacuum tube and from source to drain in a MOSFET). This allows both devices to function as an amplifier, or in some cases a switch. How electric current flows in a vacuum tube compared to a transistor is very different however. Vacuum tubes rely on a process called thermionic emission, where heating the cathode causes it to shed electrons into the surrounding vacuum. The current in transistors however comes from the drift and diffusion of electrons between the source and the drain through the solid semiconducting material that separates them.
Solid-state electronics surpassed vacuum tubes due to their lower costs, smaller size, longer lifetimes, efficiency, ruggedness, reliability, and consistency. However, when solely looking for a medium to transport charge, vacuum beats semiconductors. Electrons are able to move freely through a vacuum, where they collide with the atoms in a solid state. This process is called crystal-lattice scattering. Also, vacuums are not susceptible to the kind of radiation damage that semiconductors are, and they produce less noise and distortion than solid-state materials. When only a few vacuums were needed to operate a radio or television, their drawbacks were not that significant. However, as circuits became more complicated, it became obvious something needed to change. For example, the 1946 ENIAC computer used 17,468 vacuum tubes, weighed 27 metric tons, and took up almost 200 square meters of floor space. The transistor revolution ended these issues. The great change in electronics occurred not so much because of the intrinsic advantages of semiconductors but because engineers had the ability to mass-produce and combine transistors in integrated circuits by etching a silicon wafer with the appropriate pattern. As the technology progressed, more transistors could be put on a microchip, allowing the circuit design to become more complicated from one generation to the next.
After over 40 years, the oxide layer that insulates the gate electrode of a typical MOSFET is only a few nanometers thick, and only a few tens of nanometers separate its source and drain. While transistors can’t get much smaller, the quest for faster and more energy-efficient chips moves forward. One possible candidate to replace the traditional transistor is the vacuum-channel transistor. This combines the best aspects of the vacuum tubes and transistors and can be made just as small and inexpensively as any solid-state device. In a vacuum an electric filament is used to heat the cathode to allow it to emit electrons. Vacuum-channel transistors do not require a filament or a hot cathode. If the device is small enough the electric field across it is sufficient to draw electrons from the source by the field emission process. Removing the inefficient heating element reduces the area each devices takes up and makes the new transistor more energy efficient. Current flows in the vacuum-channel transistors would be done the same as with traditional MOSFETs, using a gate electrode that has an insulating dielectric material, such as silicon dioxide, separating it from the current channel. The dielectric insulator transfers the electric field where it’s needed while preventing the flow of current into the gate.
While the work being done with vacuum-channel transistors is in the early stages, developments could have a major impact on devices where speed is critical. The first effort to create a prototype produced a device that could operate at 460 gigahertz, approximately 10 times faster than the best silicon devices. This offers great promise for the vacuum-channel transistors to operate in the terahertz gap.