Professor Zhong Lin Wang at the Georgia Institute of Technology has devised a sensor that can harvest mechanical energy and convert it into electricity. Embedded in the boot of a soldier, for instance, the sensor could conceivably gather energy when its wearer walks and use that energy to charge batteries for a radio or flashlight, for example. Similarly, blood flow from the heart could generate energy for an implanted medical device.
At Intel, meanwhile, researchers are looking at ways to letexploit energy from RFID readers to perform additional tasks. Currently, when a reader is directed at a tag, the tag typically responds by spitting out a serial number. But by inserting a capacitor or other device that can capture energy into the tag, the stored energy could be used to power a temperature sensor or an accelerometer. If someone tried to walk off with a crate, a motion sensor could send a distress signal across the network to security. It could then be periodically recharged with a quick blast.
"You can imagine a moisture sensor. You could embed it into a building and literally never have to get at it again," said Joshua Smith of Intel Research Seattle. "Frozen foods are a big one. Security is another. Another one I heard of is blood. Blood plasma has to be kept at a certain temperature."
These devices are not perpetual motion machines, which are hypothetical machines that produce useful energy in a way that breaks the laws of physics. No such machines currently exist. But from a practical point of view, these devices come close to that ideal because they can survive on energy that otherwise would be unused and they get it on their own. Smith, in fact, calls his sensors a stab at perpetual computing.
Batteries andhave been the skunk in the sensor market for years. Futurists, scientists and others have sketched out visions in which small motes will gather data from . Sensors could also be embedded in walls to help identify intruders or, as some privacy advocates worry, spy on people.
Unfortunately, no one wants to go around replacing batteries in all these things.
Intel's WISP--or Wireless Identification and Sensing Platform--takes advantage of the dynamics of, the observation that the number of transistors on a given chip can be doubled every two years, mostly through shrinking the size of the transistors. Tinier transistors are allowing chip designers to build more intelligence into ever smaller semiconductors.
RFID chips may never get to the magical price point of 2 cents apiece, down from a range of 10 cents to more than a half-dollar. However, chipmakers can put more functionality into future chips and sell them for the same price as today's relatively "dumb" tags, Smith said.
The power requirements on low-end microprocessors are also declining because of Moore's Law.because designers are pushing the performance envelope. But the trend is going the other way on chips that run at low speeds because the electrons have to travel shorter distances.
"We're getting down to the point where the actual performance of (extremely low-end) computers is getting close to the thermodynamic limits," Smith said. "Microcontrollers have gotten so low with their energy requirements we can now power a general purpose microcontroller off an RFID reader. Compared to all power sources, an RFID reader is a relatively easy case."
One of the first WISPs that Smith developed was a chip that could tell when a lid on a box was opened.
Since then, he has developed an activity/motion sensor as well as a light sensor that can detect different levels of light (not just darkness versus light). In one experiment, a light sensor hung from a suction cup on his window delivered information about light coming through the window over a 12-hour period. However, it did so without external electricity or batteries.
The sensors devised by Georgia Tech's Wang derive their energy from zinc oxide nanowires. When the wires are bent by a probe, a negative charge is created on the side of the wire that gets stretched (the outside surface) while a positive charge builds on the compressed inside surface of the wire due to what is known as the piezoelectric effect. The charge builds at the point of contact between the probe and the wire. When the wire is relaxed, an electrical current can be detected. Although the wire continues to vibrate, current only gets discharged at the moment the strain gets relieved.
"If you continuously bent the nanowire, electricity can be generated in each cycle of the bending," Wang wrote in an e-mail. "Due to the small size, the wires are very robust. They are much tougher than bulk ceramics."
Zinc oxide works because it is both piezoelectric and a semiconductor. Other compounds also exhibit these properties, but zinc oxide has the added benefit of being nontoxic to humans, he noted.
Around 30 percent of the mechanical energy gets converted into electricity. The small size of the nanowires--which measure only 200 to 500 nanometers in length and 20 to 40 nanometers wide (a nanometer is a billionth of a meter)--means that the corresponding electrical charge produced by a single wire, or even a few million, is miniscule. Nonetheless, because of their small size they could be inserted into a variety of devices and harvest mechanical energy from walking, muscle stretching, blood pressure or even the flow of liquids. Vibration from a wire could also provide energy for these devices.
Right now, the wires are in the experimental phase. Wang next wants to build an array measuring 10 microns square containing millions of nanowires to judge how effectively energy can be generated. Wang and his group will also study ways for fabricating the wires.
Military equipment using Wang's technology could appear in three to five years, he said.