15 October 2003--A microscopic device built by researchers at Sandia National Laboratory (Albuquerque, N.M.) could lead to better photovoltaic cells, more efficient light bulbs, and the rewriting of basic physics texts.
Researchers Shawn Lin and James Fleming built a photonic lattice that emits infrared radiation only at a specific wavelength. The lattice is a type of photonic band gap crystal, in which a regular structure at the scale of microns or nanometers allows light to exist only at specific wavelengths. Because such crystals can redirect light in any direction with almost no loss, they are being eyed as material for the cladding of hollow optical fibers for very high-bandwidth communications or as devices to tune lasers and light-emitting diodes to hard-to-reach wavelengths.
With the same photolithographic methods used to manufacture computer chips, the scientists inscribed the structure they wanted in silicon. They then filled the gaps with tungsten, the same material that makes light bulb filaments, and etched away the silicon, leaving a three-dimensional waffle of tungsten rods, piled in a crisscrossing log cabin style. The size and spacing of the rods, half a micron thick and spaced 1.5 µ m apart, force the photons passing between them to fit into particular wavelengths.
When Lin and Fleming heated the device in a vacuum to 1250 °C, the typical operating temperature of a thermal photovoltaic (PV) cell, they saw a sharp emission peak at 1.5 µ m. They calculated that the peak would translate into an optical-to-electrical conversion efficiency in a PV cell of approximately 34 percent and an electrical power output of about 14 W/cm2. That's far greater than the 11 percent efficiency and 3 W/cm2 output predicted by Max Planck's Law of Blackbody Cavity Radiation.
”There was no theory predicting this was going to happen,” says Fleming ”It's just something we sort of stumbled upon, really.”
Some new physics?
Planck won the Nobel Prize in Physics in 1918 for his work describing how energy radiates from an ideal solid of a given material and varies as a function of temperature. His revolutionary insight was that the emissions could exist only at discrete energy levels, or quanta. In explaining why the observed emissions didn't match the predictions of classical physics, he laid the foundation for quantum physics--the comprehensive theory of the atom.
Despite appearances, Fleming says the work doesn't actually undermine Planck, though the scientists haven't worked out a full explanation yet. ”We believe that we're not violating Planck's law, necessarily. It's just that there's another emission process,” he says.
It appears that in their structure, the tungsten has a higher density of states than in ordinary tungsten. In other words, it may be that there are more energy levels in the material for electrons to occupy than that in the ideal solid assumed by Planck. ”You could squeeze more light into a particular band than you could normally,” Fleming says.
Shanhui Fan, an assistant professor of electrical engineering at Stanford University (Stanford, Calif.) and a photonic crystal expert, is excited about the possibilities. Rewriting Planck could potentially have implications in almost every field of optics, from sensors to switches to fibers.
”Usually when you go into a physics book, you get a standard blackbody radiation curve,” Fan says. ”They can just engineer the spectrum, which to me is really wonderful. Suddenly you have gained control over a fundamental piece of nature.”
A practical cornucopia?
The most immediate practical application could be better thermal photovoltaic cells--a concept for convert lighting to electricity, only in the infrared rather than the visible, as in current working solar cells. By squeezing more of the incoming radiation into the wavelength band the cells absorb, the photonic lattice could triple a cell's power output.
The ability to fine-tune wavelengths should also make it possible to build a light bulb that's perhaps 50 percent efficient at converting electricity to light rather than the current 5 percent, and that does it in whatever color you desire. Today's light bulbs waste most of their output as heat, but a photonic lattice bulb could squeeze the energy to desired wavelengths, so the more efficient bulb would run cooler and therefore last longer. To get the effect with visible wavelengths would require narrower spacings and smaller rods, about 100 nm thick. For a photolithography industry that's making 90-nm features on computer chips, that's not a problem, but the necessary equipment is too expensive to be readily available to Sandia researchers.
”It's very interesting, very exciting work,” says Eli Yablonovitch, a professor of electrical engineering at the University of California, Los Angeles, and one of the pioneers in the field of photonic crystals. ”It looks like it has a real chance to make thermal photovoltaics practical.”
Other possibilities include better infrared cameras and detectors, or lasers that don't need as much start-up power. Of course, 1.5 µ m is a wavelength commonly used for optical telecommunications, but Fleming doubts whether the devices could be modulated fast enough to produce a high-data-rate signal.
The researchers want to look at longer and shorter wavelengths to see what happens to the effect and figure out a theory of why it occurs. A lot of their effort will be focused on characterizing the thermal emission of the structures. The researchers chose to work with tungsten because they could use a standard process for making integrated circuits, but they'd like to try out other materials that have different material properties. ”Particular materials may be better for particular applications,” says Fleming.
”We're really trying to come up with a more comprehensive theory and see how far it extends,” he says. ”There's a lot of things to do before we really consider applications.”