Good vibrations

IMPLANTED devices, such as heart pacemakers, are a valuable part of modern medicine’s armamentarium. Their use, however, is limited by the need to renew their batteries—and this is a particular problem for those, such as cochlear implants (which improve hearing), that are inside the wearer’s head.

For obvious reasons, surgeons do not like opening heads up unless it is strictly necessary. Sometimes, therefore, the battery packs that power head implants are put in the wearer’s chest. But this means running a wire up through the patient’s neck, from the one to the other, which is scarcely satisfactory either. A way to power such implants without replacing their batteries at all would thus be welcome. And Hyuck Choo of the California Institute of Technology and his colleagues think they have one. They plan to scavenge the necessary energy from the vibrations of the vocal cords that occur when someone is talking.

Dr Choo’s power plants are small sheets of lead zirconate titanate, a substance that is piezoelectric—meaning it generates electricity when it vibrates. He knew from past work that sheets of the size he chose (just under 1cm²) resonate at around 690Hz. This is close to the F in the octave above middle C, and thus well above the normal range of the human voice. Using larger sheets would lower the resonant frequency, just as long organ pipes produce lower notes than short ones. Larger sheets, though, would be less deployable inside the body. So, instead, he sought to lower a sheet’s resonant frequency without increasing its area by carving a sinusoidal shape out of it (see picture). Such a shape must inevitably be longer than its parent rectangle’s longest sides, albeit that its length is now zigzagged. A sinusoidal sheet should thus have a lower resonant frequence than its rectangular parent.

It worked. When Dr Choo and his colleagues tested the carved sheets by exposing them to a range of frequencies and monitoring the amount of electricity generated, they found that the voltage spiked at between 100Hz and 120Hz (approximately the dominant frequencies of adult male voices), and also between 200Hz and 250Hz (the female voice’s dominant frequencies). And, although the amount of power produced is not huge, it seems adequate for the task proposed.

As Dr Choo reported on January 26th, to the International Conference on Micro Electro Mechanical Systems in Shanghai, he and his team were able to harvest a tenth of a milliwatt per square centimetre of lead zirconate titanate from the voice of a man talking at 70 decibels, which is normal speaking volume, and three-tenths from someone shouting at 100 decibels. Implants usually require a tenth of a milliwatt or less to function, so this prototype’s performance suggests a practical device might be within reach—especially as the vibrations produced by the voice travel efficiently up through the skull, meaning the generator could be integrated into an implant, rather than having to be separate from it.

Since most people are not chatterboxes, talking all the time, a practical system will still need batteries to build up charge so that the surplus can be used when needed. Intriguingly, this might even be possible when someone is asleep. Part of the sound of snoring is in the experimental device’s sweet spot. That may not be much consolation for the partners of snorers. But at least their bedmates will no longer be able to turn a deaf ear to their complaints.