MANHASSET, NY -- MIT researchers have designed a MEMS energy-harvesting device for wireless sensors that can generate two orders of magnitude more power compared to current designs.
Instead of taking a cantilever-based approach, they engineered a microchip with a small bridge-like structure that’s anchored to the chip at both ends. The Massachusetts Institute of Technology researchers deposited a single layer of piezoelectric material (PZT) to the bridge, placing a small weight in the middle of it.
When undergoing a series of vibration tests the device was able to respond not just at one specific frequency, but also at a wide range of other low frequencies.
“There are wireless sensors widely available, but there is no supportive power package,” says Sang-Gook Kim, a professor of mechanical engineering at MIT and co-author of the paper in a statement. “I think our vibrational-energy harvesters are a solution for that.”
A common energy-harvesting design consists of a small microchip with layers of PZT glued to the top of a tiny cantilever beam. As the chip is exposed to vibrations, the beam moves up and down like a wobbly diving board, bending and stressing the PZT layers. The stressed material builds up an electric charge, which can be picked up by arrays of tiny electrodes.
The cantilever-based approach has limitations and simply increasing the number of cantilever beams and PZT layers occupying a chip is wasteful, and expensive, say the researchers.
“In order to deploy millions of sensors, if the energy harvesting device is $10, it may be too costly,” says Kim. A single-layer MEMS device can be fabricated for less than $1, according to Kim.
The researchers came up with a design that increases the device’s frequency range while maximizing the power density, or energy generated per square centimeter of the chip.
When undergoing a series of vibration tests the device was able to respond not just at one specific frequency, but across a wide range of other low frequencies.
The researchers calculated that the device was able to generate 45 microW of power with just a single layer of PZT — an improvement of two orders of magnitude compared to current designs.
Nice improvement but 45uw at what cost? This should be scalable because a single layer generating X watts dont easily translate to double when the layers are doubled.
I guess a super sensitive thermocouple in Si/Ge can beat this hands down in power output.
It is interesting (and counter-intuitive) that taking two facing cantilevers and connecting them (a bridge) results in more energy being extracted rather than less. I would have thought that the connection would reduce the freedom of movement and energy recovery. Does the connection allow a much heavier weight to be supported which thereby increases the energy extraction?
It is great to hear new technology being developed for energy harvester. With more energy being generated and yet lower cost, the deployment of WSN can become more practical. The only concern of the PZT kind of energy harvester is the reliability and robustness. Any data that can prove the part to be still useful after years of operation?
I think 2W/cubic cm is pretty good improvement on the power density of the PZT energy harvesting devices. Hope to see some more progress on this space. Is there any information available on the improved frequency bandwidth?
This is a very good information. The research of this is really in a good timing now that the Internet of things and the WSN (Wireless Sensor Networks) are becoming widespread. Mobile computing will benefit a lot from using this kind of technology. Imagine... in the near future devices will recharge at the same time they´re being used.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.