The past few years have seen a quantum jump (no pun intended) in the efficiency of LEDs, so much so that they are now entering the mainstream for display backlighting units (BLUs), traffic control signals, and general purpose illumination. The increase in efficiency has come about due to advances in light output, light collection (étendue) and heat dissipation technologies; these efficiencies translate directly into lower cost and lower energy consumption. But another breakthrough is still needed, to solve a long-standing problem known as efficiency "droop."

The efficiency of electrical and mechanical devices has its optimum value at a certain point that reflects the tradeoff between useful output and the work required to achieve it, usually due to a secondary mechanism becoming dominant. Gasoline engines (and therefore automobiles), for example, have a peak efficiency somewhere in their normal operating regime, above and below which the friction of the moving parts dominates and reduces the efficiency even while total output continues to increase. This is why a national 55MPH speed limit was established in response to the gas crunch in the 70s: efficiency decreases rapidly at higher speeds and gasoline consumption increases.

White LEDs have an efficiency curve that typically peaks around 250 lm/W. If you drive the LED harder and try to pump out more light, the efficiency rapidly drops, resulting in droop. While experts believe peak efficiencies above 300 lm/W are probably just around the corner, droop is still holding back the cost-effective use of LEDs for BLUs, even though their competition, cold cathode florescent lamps (CCFLs), typically have an efficiency of only about 100 lm/W.

The cost problem comes from the fact that too many LEDs are needed in a BLU design. When you reduce droop and can drive the LEDs harder, fewer LEDs will be required to produce the same light output. Not only will the LEDs be less expensive, but fewer LED drivers will be needed.

According to a story in this month’s IEEE Spectrum, while it is known that droop affects LEDs with a conventional nitride structure, the cause of it is controversial. Some researchers believe an explanation may emerge from a better understanding of why nitride-based LEDs emit any light at all, considering the very high number of defects in a typical GaN LED junction. The article goes on to say that while Philips/Lumileds has shown some promising advances in the solution of this problem, not all industry observers are convinced of the applicability of the research.

A droop breakthrough would revolutionize an already growing market for LED BLUs. DisplaySearch recently estimated that the total number of LED-backlit LCD panels larger than 10-inch is expected to reach 104.5 million units for all of 2009. And iSuppli predicts that the proportion of TVs using LED backlights will grow from 3% in 2009 to 39% in 2013, with total shipments to reach 90 million units. DisplaySearch also believes that a 50% penetration rate will be achieved in 2014.

The widespread use of LED BLUs is closely tied to the efficiency issue. In a 14.1 inch notebook with CCFL backlight, for example, the display power consumption represents more than 30% of the entire notebook power consumption. Improve this factor and battery use improves. In a TV, the BLU can account for 2/3 of the total power consumption. When you combine lower cost and better efficiency with longer BLU life and RoHS compliancy, the penetration will continue to increase and the price gap with CCFL BLUs will narrow. With a current LED price premium over CCFL of about $150 for 40/42-inch panels, suppliers are eager to reduce this number to less than $100 for 40/42-inch panels, and to below $150 for 46-inch panels.

In a time when corporate research spending has been cut back, this represents an opportunity to generate intellectual property that may provide a huge return. The problem is no one knows how long the research will take. Any bets?