Tikalon Blog by Dev Gualtieri

Vacuum Transistors

July 11, 2014

Stating the obvious, electronic materials have the limitation that they're composed of matter. In that sense they're subject to material limitations, such as electrical conductivity when it isn't desired, and speed-limiting permittivity. As its name implies, permittivity is a measure of how a material will permit passage of an electric field. Material limitations are the reason for a revival of interest in vacuum electronic devices. I've written two previous articles on this topic, Vacuum Tube Redux (August 9, 2011), and Vacuum Transistor (May 23, 2012).

The first vacuum electronic device was the vacuum tube , a component that defined electronics through the 1950s. One indication of both the ubiquity and the lack of reliability of vacuum tubes was the presence of vacuum tube testers at the corner drug store. You could test the vacuum tubes from your television, radio, and record player, there, and buy replacements.

Vacuum tubes had a few advantages over early transistors. They were operable over a wider frequency range and at higher power. In the first few decades after the invention of the transistor, it wasn't possible to have a transistorized FM radio, or a transistorized guitar amplifier. As transistors increased in quality and declined in cost, vacuum tube technology was abandoned. There was an attempt by RCA to miniaturize vacuum tubes, but its nuvistor needed a heated cathode to generate an electron beam by thermionic emission.

Thermionic emission isn't the only way to generate electrons. The large electric field at the tip of a pointed conductor will produce electrons by field electron emission. In theory, the electric field there is infinity, which would definitely remove electrons from a conductor. In practice, the electric field, as defined by the Fowler–Nordheim equations, is large, but less. The importance of these equations by British physicist, Ralph H. Fowler and German-American physicist, Lothar Nordheim was underscored to me by the phenomenal number of times they were mentioned by the professor of one of my graduate courses.

In the 1970s, Charles Spindt and his colleagues at the Stanford Research Institute (Menlo Park, California) fabricated miniaturized field emission vacuum tubes by the same planar processing techniques used for integrated circuits .[1-2] These devices, as illustrated in the figure, also worked at a lower voltage than thermionic vacuum tubes because of their small dimension.

Fig. 1 of US Patent No. 3,755,704, "Field Emission Cathode Structures And Devices Utilizing Such Structures," by Charles A. Spindt, Kenneth R. Shouldersand Louis N. Heynick, August 28, 1973.

(Via Google Patents).[2]

The electrical characteristics of the Spindt architecture were somewhat difficult to control, but in 2011, a research team from Vanderbilt University made a field emission device from diamond , as shown in the figure.[3-4] Production of diamond layers is presently fairly routine using the reaction of a hydrogen with methane . Diamond has an extremely low work function , so it will produce field emission electron beams at low voltage.[5]

Left, electron micrograph of a nanodiamond field emission device. Right, detail of the device, showing the nanodiamond field emitter cantilevered above the insulating silicon dioxide surface. The silicon dioxide layer is on a silicon wafer . (Images, Davidson Lab, Vanderbilt University).[3]

The diamond devices are operable from liquid nitrogen temperature (77 K) all the way to about 500°C. They would also be inherently radiation resistant , so they would be useful in satellites and spacecraft. Two problems with this architecture are the definition of the gate , and the fact that some current is drawn through the gate.

Scientists from the NASA Ames Research Center (Moffett Field, CA) developed a planar process electron emission device in 2012 that they call a "vacuum channel transistor."[6] They built on previous work by the National Nanofab Center (Daejeon, Korea).[7-8] The device, as shown in the figure, uses a vertical architecture like the geometry of a standard MOSFET .

NASA Ames Research Center vacuum transistors

The 2012 (left) and 2014 (right) versions of the NASA Ames Research Center vacuum channel transistor. (Illustration by the author using Inkscape.)

The spacing between the electron emitter and collector can be made as small as 150 nm, which allows operation at less than 10 volts. These devices functioned to 460 GHz , and operation at terahertz range is predicted.[6]

The mean free path of electrons in air is longer than the gap, so this vacuum device doesn't really need a vacuum. If the gas, helium, is used, instead of air, the mean free path is 1000 nm. Since the velocity of electrons in semiconductors is only as high as 5 x 10⁷ cm/sec, while it's 3 x 10¹⁰ cm /sec in a vacuum, operation at high frequencies is assured.

The Ames team has updated their architecture by bottom gating, also shown in the figure.[9-10] Although the Ames vacuum channel transistor operates at ten volts, which is much higher than the 1.2-1.6 volt range of the present generation of MOSFETs, operation at lower voltage has been reported by others.[9] Sharper points might help, but there might be at a detriment to lifetime.[9]

The Ames team next plans to demonstrate the first vacuum channel transistor integrated circuit.[9] There's still some competition from semiconductor transistors. Operation of silicon-germanium transistors has been demonstrated at 417 GHz at room temperature.[10] As I've repeatedly written, many scientific discoveries happen by chance. The Ames research started with an experiment to oxidize a nanowire. The oxidation divided the wire into two pieces, and it was realized that such nanogaps might be useful.[9]

Vacuum Transistors

References: