Piezoelectric Antennas For Very, Very Low Frequencies

If you want to talk about antennas, the amateur radio community has you covered, with one glaring exception. Very low frequency and Extremely Low Frequency radio isn’t practiced very much, ultimately because it’s impractical and you simply can’t transmit much information when your carrier frequency is measured in tens of Hertz. There is more information on Extremely Low Frequency radio in Michael Crichton’s Sphere than there is in the normal parts of the Internet. Now there might be an easier way to play with VLF radiation, thanks to developers at the National Accelerator Laboratory. They’ve developed a piezoelectric transmitter for very long wavelengths.

Instead of pushing pixies through an antenna, this antenna uses a rod-shaped crystal of lithium niobate, a piezoelectric material. An AC voltage is applied to the rod makes it vibrate, and this triggers an oscillating electric current flow that’s emitted as VLF radiation. The key is that it’s these soundwaves bouncing around that define the resonant frequency, and the speed of sound in lithium niobate is a lot slower than the speed of light, but they’re translated into electric signals because of its piezoelectricity. For contrast, if this were a wire quarter-wave antenna it would be tens of kilometers long.

The application for this sort of antenna is ideally for where regular radio doesn’t work. Radio doesn’t work underwater, but nuclear subs trail an antenna out of the back to receive messages using Extremely Low Frequency radio. A walkie talkie doesn’t work in a mine, and this could potentially be used there. There is a patent for this piezoelectric antenna, so if anyone knows of a source of lithium niobate, put a link in the comments.

We’ve seen this trick before to make small antennas even smaller, but this is the first time we’ve seen it used in the VLF band, where it’s arguably even more impressive.

40 thoughts on “Piezoelectric Antennas For Very, Very Low Frequencies

    1. It’s not about efficiency: “The key is that it’s these soundwaves bouncing around that define the resonant frequency, and the speed of sound in lithium niobate is a lot slower than the speed of light, but they’re translated into electric signals because of its piezoelectricity. For contrast, if this were a wire quarter-wave antenna it would be tens of kilometers long.”

      1. That’s not different from any piezoelectric resonator. Its resonance is mechanical in nature, and it is excited by, and produces, electricity.

        There isn’t really a point in making an antenna of it unless it imore efficiently couples the input power to far-field radiation than a wire coil of the same dimensions. A “screwdriver” antenna, a whip brought to resonance with a coil that has a motorized moving tap, is only a few percent efficient at 1-3 MHz (and the coil, rather than the whip, may do most of the radiation). That’s what this would need to beat.

        1. Down in VLF, any portable antenna is a Hertzian dipole. Trying to make it radiate a meaningful amount of power is a matter of matching a tiny radiation resistance to a power amplifier’s output impedance. That means a resonant network of some kind, and L-C networks have rather low Q factors compared to the impedance ratio you’re trying to match, thanks mostly to the resistance of the inductor. That, in turn, means that most of your power is burned off in the losses of the resonator instead of being radiated.

          Piezoelectric resonators are well known for having better Q factors than L-C resonators — it’s why we use them for frequency control. The bandwidth of a resonator is inversely proportional to its Q factor, so you get better frequency stability out of a piezoelectric crystal than you do out of an L-C network.

          Using the mechanical resonance of the rod as the loading network gets you a better Q than a coil, and the piezoelectric effect allows the rod to serve both as resonator and antenna. It’s an awfully small antenna, and coupling the same resonator to a longer antenna would probably help, but a longer antenna might not be an option if the goal is a portable system.

          1. The size of a resonant dipole is reduced by the ratio of speed of light to speed of sound in the crystal. So this invention makes it practical to use resonant dipoles in the VLF band. They’ve achieved a Q of 12,000 and hope for 100,000. This looks interesting!

          2. A high Q does mean that the bandwidth available for data transmission is incredibly small. Even with the conventional huge VLF antennas like the one from SAQ this was already a problem.

    2. Nature has a better article: https://www.nature.com/articles/s41467-019-09680-2 It’s all about the high-Q, without lots of impedance matching circuitry, that allows the device to couple energy to the far field over time. They then go on to say that of course high-Q usually implies very low modulation bandwidth, as it takes a lot of time to change the frequency of the signals bouncing up and down many times.. however this device mechanically alters it’s resonance, which has immediate effect on radiated frequency and can thus be modulated at a higher rate/bandwidth.

    3. It’s about resonance. A thin wire would resonate, but would be very narrowband. If you slow down the wave, you can make the antenna smaller. A clever way to do that is by turning it into an acoustic wave, which is much slower than an EM wave. It also makes it a little less efficient than a long wire, but that’s good because it broadens out the bandwidth a little bit and that’s the only way you can send information.

      The same principle is used in bulk acoustic wave filters, which shrink them down immensely. Acoustic delay lines have been used (before those newfangled digital delays) to make delays without using long cables.

      The resultant antenna is nowhere near as efficient as a dipole, and the gain is low, essentially a point source. Which is OK, an omni radiator doesn’t need to be aimed.

      Loopstick antennas in AM radios are terribly inefficient, but resonant, so they work.

      Public relations people are excellent at playing ‘telephone’ with researchers, and turning around to explain (in their own words) to the public how some piece of research is going to revolutionize something or other…. after further funding. The popular press reinterprets it and publishes it in their own words… so it goes.

  1. There are lithium niobate rectangles about 0.9 inch long on sale on ebay. 3 for $33. Perhaps they could be tested with a higher frequency, like HF radio.
    Hey, I’m no physicist. But my quartz crystals are not at all efficient in coupling their resonance to the far field. It’s confusing to me that these would be.

    1. I admit the same without doing more homework. So what if things move slow in the antenna? Seems the speed of light outside the antenna has to somehow be matched to too. It’s a rather extraordinary claim, and if it’s just the speed in the media, it could have been duplicated long ago by a tiny phased array driven by a tapped delay line…
      Some of the verbiage sounds like someone thinks there’s a way to cheat Maxwell’s equations. Of course, ridiculous errors are par for the course in all science reporting, due to some editor who doesn’t really understand himself dumbing what he thinks down to what he thinks the audience will understand = better at communicating, but wrong. Or the other case of maximizing grant income possibilities via “press release science”. Could be neither as well – this might actually work – needs homework.

    2. Your quartz crystals are not efficient radiators for two reasons:
      1) They are shielded.
      2) They are configured as thin wafers, thus the distance between the electrodes is very small.

      1. Sid,

        Yes, the crystals in my modern ham radio are that way. But vintage ham crystals were mounted in phenolic cases, and hams regularly modified their crystals before there were stable VFOs available to them. It was not unusual for a ham to file a crystal in order to move its frequency, and to do all sorts of fooling around with overtones. And of course the old crystals were thicker. You can still find them at ham flea markets.

        I’m really dubious about this – the fact that someone can receive a signal a few feet away doesn’t make them _good_ antennas, or indeed any better than a coil of wire.

          1. I’m still wondering why a century of empiricists working on antennas would not have come up with this. Perhaps later on they will publish better figures and perhaps demonstrate a successful antenna before others.

          2. I suppose the empiricists who you are thinking of have restricted their investigations to various configurations of copper wire. Allowing for exotic materials, Magnetoelectric antennas, and/or piezoelectric antennas (with sizes as small as one-thousandth of a wavelength in air) demonstrate 1–2 orders of magnitude miniaturization over state-of-the-art compact [copper wire] antennas without performance degradation. Read about ME antennas here: https://www.nature.com/articles/s41467-017-00343-8

          3. The Nature article gives an electromagnetic coupling coefficient of 1.36%. This says the antenna is resonant, but that the vast majority of power goes to heat. It might still be effective as a receiving antenna – but I think that we’re waiting for someone to achieve 10% or so before this would be desirable for transmit, to be competitive with shortened wire designs.

          4. I’ll accept that a dielectric antenna, if it can be made to handle the power, could be an improvement over a screwdriver at 80 and 160, with the quoted 1.36% efficiency. Although those figures are for a whip tuned to resonance with, I presume, the coil not acting as a radiator. In the case of the screwdriver antenna, the coil, rather than the whip, may be the main radiator. This doesn’t say it’s efficient.

      1. I get the feeling this works more along the lines of a lightbulb in that it puts out light but isn’t useful for receiving it. If it does work as a receiver at all, the reception must be terrible, probably to the point that the signal is buried in the noise floor.

  2. Finally I can send the launch codes to my Arduino controlled mini ballistic missile submarine without swimming out and mumbling radio noises as I push the launch button by hand!

    1. Too low data rate. ELF is normally only used to send a short code to a submarine, signalling them that command wishes to communicate.

      They then go up high enough to release a sat communication buoy to the surface.

    1. If the blurb is accurate when it says that nuclear subs still trail an antenna behind them, we can assume that either (A) the military didn’t know about this, (B) the military knew and tried it but it wasn’t effective enough to replace their current antenna, or (C) the military knew and tried it and found it to be effective, but it’s still working its way up the approvals and manufacturing chain and hasn’t yet made it to active use.

      Personally, my money’s on (C).

  3. I’m a ham, and very interested in a much higher, but still really freaking low, band: 742mhz, or 630m. My question would be whether or not either a commercial product or a good design to build your own antenna will be available for higher bands. For my purposes a 615m dipole is just as impractical as one kilometers long. It sounds like these antennas are more technically challenging than the coil loaded Marconi or magnetically coupled ones people are building now.

  4. If anyone is interested in purchasing large stoichiometric lithium niobate or lithium tantalate single crystals (50mm diameter, 10-50mm thickness in z-direction), contact me at ikeepcommas (at) gmail (dot) com

    You can also use these to make super-compact x-ray and neutron sources (see nature articles on pyroelectric fusion)

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.