• A variety of silicon chip micro-reactors developed by the MIT team. Each of these contains photonic crystals on both flat faces, with external tubes for injecting fuel and air and ejecting waste products. Inside the chip, the fuel and air react to heat up the photonic crystals. In use, these reactors would have a photovoltaic cell mounted against each face, with a tiny gap between, to convert the emitted wavelengths of light to electricity.

    Photo: Justin Knight

    Full Screen
  • From left, research engineer Ivan Celanovic ScD '06, graduate students YiXiang Yeng and Walker Chan, and Professor of Physics Marin Soljačić.

    Photo: Justin Knight

    Full Screen

Sun-free photovoltaics

Materials engineered to give off precisely tuned wavelengths of light when heated are key to new high-efficiency generating system.

A new photovoltaic energy-conversion system developed at MIT can be powered solely by heat, generating electricity with no sunlight at all. While the principle involved is not new, a novel way of engineering the surface of a material to convert heat into precisely tuned wavelengths of light — selected to match the wavelengths that photovoltaic cells can best convert to electricity — makes the new system much more efficient than previous versions.

The key to this fine-tuned light emission, described in the journal Physical Review A, lies in a material with billions of nanoscale pits etched on its surface. When the material absorbs heat — whether from the sun, a hydrocarbon fuel, a decaying radioisotope or any other source — the pitted surface radiates energy primarily at these carefully chosen wavelengths.

Based on that technology, MIT researchers have made a button-sized power generator fueled by butane that can run three times longer than a lithium-ion battery of the same weight; the device can then be recharged instantly, just by snapping in a tiny cartridge of fresh fuel. Another device, powered by a radioisotope that steadily produces heat from radioactive decay, could generate electricity for 30 years without refueling or servicing — an ideal source of electricity for spacecraft headed on long missions away from the sun.

According to the U.S. Energy Information Administration, 92 percent of all the energy we use involves converting heat into mechanical energy, and then often into electricity — such as using fuel to boil water to turn a turbine, which is attached to a generator. But today's mechanical systems have relatively low efficiency, and can't be scaled down to the small sizes needed for devices such as sensors, smartphones or medical monitors.

"Being able to convert heat from various sources into electricity without moving parts would bring huge benefits," says Ivan Celanovic ScD '06, research engineer in MIT's Institute for Soldier Nanotechnologies (ISN), "especially if we could do it efficiently, relatively inexpensively and on a small scale."

It has long been known that photovoltaic (PV) cells needn't always run on sunlight. Half a century ago, researchers developed thermophotovoltaics (TPV), which couple a PV cell with any source of heat: A burning hydrocarbon, for example, heats up a material called the thermal emitter, which radiates heat and light onto the PV diode, generating electricity. The thermal emitter's radiation includes far more infrared wavelengths than occur in the solar spectrum, and "low band-gap" PV materials invented less than a decade ago can absorb more of that infrared radiation than standard silicon PVs can. But much of the heat is still wasted, so efficiencies remain relatively low.

An ideal match

The solution, Celanovic says, is to design a thermal emitter that radiates only the wavelengths that the PV diode can absorb and convert into electricity, while suppressing other wavelengths. "But how do we find a material that has this magical property of emitting only at the wavelengths that we want?" asks Marin Soljačić, professor of physics and ISN researcher. The answer: Make a photonic crystal by taking a sample of material and create some nanoscale features on its surface — say, a regularly repeating pattern of holes or ridges — so light propagates through the sample in a dramatically different way.

"By choosing how we design the nanostructure, we can create materials that have novel optical properties," Soljačić says. "This gives us the ability to control and manipulate the behavior of light."

The team — which also includes Peter Bermel, research scientist in the Research Laboratory for Electronics (RLE); Peter Fisher, professor of physics; and Michael Ghebrebrhan, a postdoc in RLE — used a slab of tungsten, engineering billions of tiny pits on its surface. When the slab heats up, it generates bright light with an altered emission spectrum because each pit acts as a resonator, capable of giving off radiation at only certain wavelengths.

This powerful approach — co-developed by John D. Joannopoulos, the Francis Wright Davis Professor of Physics and ISN director, and others — has been widely used to improve lasers, light-emitting diodes and even optical fibers. The MIT team, supported in part by a seed grant from the MIT Energy Initiative, is now working with collaborators at MIT and elsewhere to use it to create several novel electricity-generating devices.

Mike Waits, an electronics engineer at the Army Research Laboratory in Adelphi, Md., who was not involved in this work, says this approach to producing miniature power supplies could lead to lighter portable electronics, which is "critical for the soldier to lighten his load. It not only reduces his burden, but also reduces the logistics chain" to deliver those devices to the field. "There are a lot of lives at stake," he says, "so if you can make the power sources more efficient, it could be a great benefit."

The button-like device that uses hydrocarbon fuels such as butane or propane as its heat source — known as a micro-TPV power generator — has at its heart a "micro-reactor" designed by Klavs Jensen, the Warren K. Lewis Professor of Chemical Engineering, and fabricated in the Microsystems Technology Laboratories. While the device achieves a fuel-to-electricity conversion efficiency three times greater than that of a lithium-ion battery of the same size and weight, Celanovic is confident that with further work his team can triple the current energy density. "At that point, our TPV generator could power your smartphone for a whole week without being recharged," he says.

Celanovic and Soljačić stress that building practical systems requires integrating many technologies and fields of expertise. "It's a really multidisciplinary effort," Celanovic says. "And it's a neat example of how fundamental research in materials can result in new performance that enables a whole spectrum of applications for efficient energy conversion."

David L. Chandler contributed to this story.

Topics: Energy, MIT Energy Initiative (MITEI), Solar


the heat is converted into the precise light wavelengths that photovoltaic cells can best convert to electricity... is the converse possible ? Can the precise wavelengths of light can be converted to heat to achieve higher thermal efficiency ...it would be quite useful in heating systems i think
Wouldn't it be more efficient to tune the photonic crystal to emit microwaves instead of light, and then use a rectenna (array) to convert to electricity?
Excellent research by MIT. Yes. I found long back when I was in New York in Christmas Season wearing Photosun glasses in the midst of snow that my Glasses turned black. It is because of radiation rather than direct sunlight. By using heat at outsdide Box Type Air conditioners(Which are in use in millions) one can generate Power through PV. Yet another great innovation from MIT Researchers. Congratulations. Dr.A.Jagadeesh Nellore(AP),India E-mail: anumakonda.jagadeesh@gmail.com
Could use electricity to power electric fan mounted on heat sink(CPU for example) - if it is hot - fan is activated if it is not - no current produced...
Correct me if I get this wrong. When the tungsten gt heated up, there should be a broadband emission and the photonic crystals are used to "select" certain wavelengths for resonance.This is like laser cavity resonance, using tungsten as gain medium and photonic crystals as resonantors. There are only certain wavelengths emitted, which means the rest (the majority)of wavelengths are still wasted, right? Is it possible this kind device being efficient?
If this is in a warm closed system, is it still able to get current (and hence work)?
Could this technology be used to make high-efficiency light bulbs? If you devide the glow wire into 3 or more sections, each tuned to another color of light, you could produce RGB-white or a more complex spectrum. If I get it right, the infrared emissions are reduced dramatically. So the efficiency should highly increase. Even if efficiency and duration of life are smaller than that of LEDs, and if costs are higher, there could be an advantage: you have to cool power LEDs, whereas you don't even want to cool light bulbs of higher power. And for red light bulbs you could reduce temperature and increase durability of the tungsten wire.
It can't work in *warm* systems. You need *hot* systems. In a closed system of 3000 K the photovoltaic cells will be destroyed. If they could stand the heat, then you wouldn't need this technology. The PV cells ought to work with the normal thermal visible radiation of themselves. But under those hot conditions there would be no effective band gap, so conversion to electric energy would be impossible.
"While the device achieves a fuel-to-electricity conversion efficiency three times greater than that of a lithium-ion battery of the same size and weight" First let me say that in itself is outstanding; however, when can we expect a release of the data? Lithium ion batteries cover a somewhat broad range of efficiency ratings based on when and how they were manufactured.
Start to sell these cells on ebay right now and spread the word before they'll begin to "eliminate" this project from the earth's surface. One step and then another to develop this technology in other ways... Just produce some cells ajusted to standard PV cell dimensions and give them away to the people so they can use them already and do self projects if this works as you say. Nice job in deed.
I work in the solar industry and I am constantly amazed at the ultra- conservative estimates that industry pundits make in terms of the improvements in efficiencies that are anticipated in solar cells. These kinds of breakthrough technologies will get to market and change the industry overnight. Despite the best efforts of the oil and coal industries, these technologies will emerge in a big way over the next decade.
What is correct?
the efficiency would then again depend on that of the array and the conversion. This is more efficient and still offers a cheap setup.
When will this technology be available??? 2014? rather 2020? I'd buy loads of them (if the costs are ok)
Back to the top