Lightimes Online News and Resources for the LED and Solid State Lighting Industry and its Supply Chain

January 19, 2012...I read a comment recently that suggested we, the industry, would need some big (significant, fundamental) breakthroughs in order to achieve the levels of efficiency that were needed to realize the ultimate success of LED lighting. While "breakthroughs" are always handy in the technology world, they usually don't need the modifier "big" in front of them to create a success. In an oft-repeated persistence meme, "A big shot is just someone who took little shots, but just kept shooting." So it goes with LEDs and LED lighting. Success won't be evading us even if the material science topped out right where it is right now (which it won't), or even if other conspiracies forced an end to the core efficiency improvements. While GM and "big oil" are said to have quashed the 200 mile-per-gallon carburetor, it appears the analogous suspects in light bulbs (GE, Philips, et al), have failed miserably to stop LED technology in time to stave off the eventual demise of incandescent and fluorescent lamps (... although they are probably sitting on alien lighting technology that delivers millions of lumens per nano-watt, but that's a different story). That LEDs will succeed is good news, though, as it means we'll have time to circle back around to the R&D that will power our cars on di-hydrogen oxide (also known as H20).

In defense of the proposition that we're going to get there regardless, its worth a quick word on the nature of lumens, since "lumens per watt" is typical recognized as "the goal" for lighting technology improvements. Not to get too "sciency" here (especially since I can't bluff that well), it's important to understand that all lumens are not created equally, in the sense that differing "optical power" is needed to create differing color lumens. "Optical power" which is the measure of the energy represented in the photons that something, an LED in this case, emits. Lumens are about the "perceived" brightness as interpreted by the human eye which, photopically speaking, peaks at about 555 nm, which is that green-yellow transition area. Not surprisingly, our eyes balance out the different frequencies of light to help us see the world the way we're "supposed to" (a self-defining kind of thing). The natural starting point for all this is the sun, since that's been main light source for humans, and the rest of the earth, for more than a couple of years now. The big "for instance" is our sensitivity to that yellow-green light. The sun generates a nicely shaped spectral "energy" curve that starts out at about 1 W/sq.meter/nm in the deep blue, peaks about 40% above that in the blue-to-green range, and then tapers back down to most of the way back to around 1.25 W/sq.meter/nm after the red (with some variance in those numbers based upon altitude and atmospheric conditions). There is a noticeable down-blip in along the green-yellow border, conveniently there in the area that challenges us, maybe helping us to see things we need to find, such as food, which is often found in the midst of green leaves (fruit) or is green itself (veggies). Maybe that explains why lots of people prefer green vegetables but don't like beets... not in the peak color sensitivity.

Why that all matters, is to make sure we rule out any "gaming the lumens" approach to efficiency gains. One simplistic cheat to more lumens per watt would be to pump up the green, and the other colors in the "center" of the spectrum that we're more sensitive too, while dialing back the colors we're less sensitive too. Anytime you've seen a LED that had a greenish tinge to it, you've experienced that skew, whether it was intentional or not. Interesting to note is that native green LEDs are the least efficient kinds, so it's fortunate that we don't need much green in order to have accurate color perception... seems an interesting coincidence. I suspect a few manufacturers in the earlier days may have had runs with a new phosphor approach that turned out a bit green, and while not planning to carry that over to the volume product, it could have come in handy for quoting the leap forward in total lumens the "white" LED was producing. CRI and eventually CQI, the more accurate and technology independent Color Quality Index, will help keep this in check as new players of significance appear on the LED manufacturing scene. And that can be expected, as China's newest 5-year plan has called for those many small LED manufacturers who enjoyed MOCVD equipment subsidies, to please now merge themselves into just a few new LED manufacturing powerhouses. It should be noted that blue-tinged white LEDs similarly offer more lumens-per-watt, but for different reasons. In that instance, it's simply that less of the original blue photons are converted by the phosphor, and come right on through unmolested, able to register lots of lumens in our blue-sensitive eyes. Conveniently we can apply the existing lighting term of "cool" to the resulting color family, so it's not a fault, just a feature.

With laboratory results of white LEDs having been announced at over 200 lumens per watt, and production devices now at 150, it would present a compelling case that we'll get above 200 for production LEDs in the near future. If we assume that's as far as the efficiency game will take us, is it good enough? I think so, for two simple reasons: 1) These are semiconductor devices, and if there is one thing the last 30 or so years have taught us, semiconductor manufacturing know-how continues to march forward and one inevitable result is decreasing costs. Whether from economies of scale, process tweaks or progressive innovations, such as Bridgelux' and Osram's confident announcements of GaN-on-silicon, the cost per lumen will continue to fall; and 2) We can now fit enough lumens in a square inch to replace every incumbent lighting technology in every application that exists today.

Size does matter, and I can remember clearly when cool-white LEDs crossed the lumens-per-square-inch threshold that enabled the launch of LED-based streetlights in a standard cobra-head envelope. The thing on the end of existing poles needed to be no bigger than what was there already, for a host of reasons, and solid state lighting had finally pulled it off. Much the same as when LEDs proved themselves in traffic signals, the transition became inevitable, an in just a few years, we'll be smirking when we see an "antique" non-LED cobra-head lighting up a roadway.

LED manufacturers have stayed dead on the line to the target in continuing to shrink LED footprints to enable that "holy grail" of lighting, the replacement lamp market. Lumileds fired that first round some years ago with the Luxeon Rebel family, which is currently packaging up to 200-ish warm lumens into a 3mm x 4.5mm package (which could be referred to as a "3045") when driven at 1000mA yielding around 80 lumens per watt. The tighter spacing provided enough lumens per square inch to better tap some of the more challenging lamps, such as the MR-16 format. Most recently, Cree announced its XB-D which shrink things still further into a 2525 package, allowing around 12 200+ lumen devices to be squeezed into the space that would normally fit 7-8. (Both numbers are referencing high drive currents at around 3500K... they can be driven at 350mA, with efficacies approaching 130 lm/w). While more lumens per chip is generally the thing we think of in terms of packaging efficiency, being able to fit more individual chips in a given piece of real estate also has advantages in terms of scalability, and the option to driving them "more gently" for better efficiency and/or longer lifetimes. Neither is "the best" in a blanket sense, as there are always design tradeoffs and application considerations where big chips make sense, but the upshot is that we can fit the needed number of lumens pretty much wherever we need to, and cost optimization has come right along with that.

While LEDs "taking over the world" is exciting in itself, we're always looking for the next thrilling development, and there plenty of reasons to expect that 200 lm/w isn't the end of the road. Nor is a 2525 device as compact as we expect a lighting-capable LEDs to get. The incumbent technologies are no-doubt doomed and those that do doubt will find themselves left behind when "suddenly" LEDs are the only answer, and they're still wondering if they'll really last 50,000 hours (despite passing a traffic signal on their daily commute and not realizing it contains the same set of LEDs that were installed in it 100,000 hours ago). No, we'll see more efficiency, and lower costs, with the result being innovative new ways to package and deliver light, as well as whole set of new ways to use light that we hadn't really ever dreamed of before.

Join the movers and shakers that are helping to define and shape the future of smarter, better lighting at the next installment of the SSL Summit, coming up April 3-4 in Long Beach, California.

TE Connectivity Introduces Socket for Nichia LEDs
LIGHTimes News Staff

Opto Diode Introduces High-Power, Narrow-Angle IR LEDs
LIGHTimes News Staff

New Pixled Displays to be Launched at ISE Europe
LIGHTimes News Staff

Active-Semi Announces New LED Driver Technology Which Company Claims Reduces Component Count and Lasts Longer
LIGHTimes News Staff

Avago Technologies Enables Thinner Digital Cameras with New Small-Footprint Auto-Focus Auxiliary Flash LEDs
LIGHTimes News Staff

Mitsubishi Electric Diamond Vision Systems Wins an Engineering Emmy
LIGHTimes News Staff

Eulitha to Collaborate with EV Group to Develop To Develop Low Cost of Ownership Nanopatterning Solution for HB LED Fabrication
LIGHTimes News Staff

Optics for Hire Releases First Lens Design Customized for Cree XB-D
LIGHTimes News Staff

Squeezing Out More Lumens for Less
Tom Griffiths - Publisher

TE Connectivity Introduces Socket for Nichia LEDsLIGHTimes News Staff

Opto Diode Introduces High-Power, Narrow-Angle IR LEDsLIGHTimes News Staff

New Pixled Displays to be Launched at ISE EuropeLIGHTimes News Staff

Active-Semi Announces New LED Driver Technology Which Company Claims Reduces Component Count and Lasts LongerLIGHTimes News Staff

Avago Technologies Enables Thinner Digital Cameras with New Small-Footprint Auto-Focus Auxiliary Flash LEDsLIGHTimes News Staff

Mitsubishi Electric Diamond Vision Systems Wins an Engineering EmmyLIGHTimes News Staff

Eulitha to Collaborate with EV Group to Develop To Develop Low Cost of Ownership Nanopatterning Solution for HB LED Fabrication LIGHTimes News Staff

Optics for Hire Releases First Lens Design Customized for Cree XB-DLIGHTimes News Staff

Squeezing Out More Lumens for LessTom Griffiths - Publisher