Light-emitting diode (LED) lamps are becoming more common these days, and their prices are dropping from the near-stratospheric $60 or so you had to pay for a bulb that'd fit in a standard socket four years ago. But LEDs get less efficient the more current is run through them -- the physicists call this efficiency loss "droop" -- and inefficiency from droop is an obstacle to wider LED use. A recent study by researchers at UC Santa Barbara and France's École Polytechnique has uncovered the reason for the "droop," raising hopes of circumventing the problem.
The "droop" problem has kept LEDs almost completely restricted to applications in low-current electronics, though there are an increasing number of bulbs available for standard sockets, and they do put out a pretty good amount of light, as indicated by this quick and dirty video ReWire just captured of our 60 watt-equivalent office LED:
But 120 volt household circuits carry a lot more current than the wiring in your tablet computer, and so the amount of light per watt ReWire's LED puts out is much lower than the equivalent figure for that tablet. As a result, room lighting takes a whole lot more LED than would be the case if not for droop: to make up for the lower output per watt, you need a lot more Diode to Emit Light.
If we could overcome the droop problem, LEDs would become far more efficient and significantly cheaper, putting out three times as much light per watt as comparable compact fluorescents, and cutting our global energy consumption way down.
Finding out what causes droop is more than half the battle, and it looks as though UC Santa Barbara's Center for Energy Efficient Materials has pinned the blame on a process called Auger Recombination.
At low current, LEDs produce light by sending that current between two semiconductors. One of the semiconductors emits free electrons; the other produces what physicists call "electron holes" -- places in an atom's electron shell where an electron could fit stably according to the rules that govern the interactions of electrons and atomic nuclei, but which don't actually have electrons in them at the moment. When an electron travels from one semiconductor to the other, it falls into an electron hole. In doing so it gives up some energy, which radiates away in the form of a photon -- a particle of light.
That's at lower currents. Higher currents pump more energy into the system, generating more electrons from the first semiconductor. But increasing the number of electrons means there's a greater chance that the electrons will bump into one another rather than falling into waiting electron holes. When the electrons do the bump they lose some of their energy to heat, and some of the electrons leave the system and never find their holes. That means less energy to release as light. Those electrons that siphon off the first electron's energy are called "Auger electrons" after one of the earliest physicists to observe them, though they should really be called "Meitner electrons" after their actual discoverer, Lise Meitner. Sexism in the STEM fields: who'da thunk it?
Anyway, that interaction -- called "Auger recombination," which would also be an excellent title for a Robert Ludlum novel -- turns out to be the culprit behind LED droop, as a team led by UCSB researchers James Speck and Claude Weisbuch -- the latter of whom is also on the faculty at École Polytechnique -- developed a way to measure the energy signatures of electrons being emitted from a standard LED, and the results were consistent with Auger recombination. Others had speculated that Auger recombination might be the reason for LED droop; this research would seem to confirm that speculation.
Now that we've got a culprit behind the droop problem, the world's graduate students can get to work to find a way to design around it. If droopless LEDs do hit the market sometime soon, the benefits could be substantial. In theory, LEDs could put out 300 lumens for each watt of electrical power they're fed. The US Department of Energy has set an R&D goal of a 224 lumens per watt LED by 2025, which would be more than three times as thrifty as compact fluorescents (which run about 75 lumens per watt) and completely outshine standard incandescent bulbs (at a measly 15 lumens per watt.)
A report done by Navigant Consulting for the DOE in January 2012 offered an astonding estimate of the potential benefits of LED lighting:
In 2030, the annual energy savings due to the increased market penetration of LED lighting is estimated to be approximately 300 terawatt-hours, or the equivalent annual electrical output of about fifty 1,000-megawatt power plants. At today's energy prices, that would equate to approximately $30 billion in energy savings in 2030 alone. Assuming the current mix of generating power stations, these energy savings would reduce greenhouse gas emissions by 210 million metric tons of carbon. The total electricity consumption for lighting would decrease by roughly 46 percent relative to a scenario with no additional penetration of LED lighting in the market--enough electricity to completely power nearly 24 million homes in the U.S. today.
For "fifty 1,000 megawatt power plants," you can read "22 coal-fired plants the size of the Navajo Generating Station" or "21 San Onofre nuclear power plants." Sounds good to us.