Electrical engineers who attended school during the early 1970s may be able to recall their first exposure to light-emitting diodes (LEDs). It was often in a circuits lab where the instructor wired up a single LED on a breadboard — mere undergrads were generally not allowed anywhere near the novel and expensive components. Students huddled around the instructor were astounded to see a dim, red light coming from a tiny spot on the circuit board.
Thus were the humble beginnings of solid-state lighting, the vast majority of it comprised of LEDs. Today LEDs continue to get brighter, but the main trends in place for solid-state illumination include a quest for reduced manufacturing costs and initiatives to make over-enthusiastic LED lighting suppliers meet their own published specifications.
It isn’t just developed countries adopting LED illumination. China recently installed more than 10,000 street lights hosting about a million Cree high efficiency LEDs on a highway project in Shenzhen. And like the U.S. and Europe, China is phasing out incandescent bulbs. It is banning the sale of 100-W incandescents starting next October. Sixty watters will get the axe there two years later, 15-W bulbs two years after that.
The market research firm DisplaySearch says lighting installations like those on the Shenzhen highway help explain why the market for LEDs was $7.2 billion last year and is expected to hit $12.7 billion by 2014. China is the biggest market for LED street lighting, thanks to policies of the Chinese government toward infrastructure investment and energy efficiency.
Another market research company, IMS Research, believes LEDs will account for 40% of the total global lamp market by 2013, thanks to longer lifetimes and better energy efficiency. Efficiency is a key in that experts say nearly 20% of the world’s energy resources go toward illumination. The U.S. Dept. of Energy (DoE) estimates that use of solid-state lighting could save between 10.5 and 16.0 quads of energy over the next 20 years. (An energy quad is equal to the average annual per-capita energy consumption of 2.9 million people, or about 16 million barrels of oil.) And these figures do not include energy and cost savings obtained from less maintenance.
Illumination-grade LEDs are still expensive but their cost is dropping every day. GE Lighting Solutions president Jaime Irick says LED costs have been dropping about 20% annually and thinks the trend will continue. But he figures it will be about 10 years before LED lighting will be able to compete on price with other energy efficient lighting technologies, such as compact fluorescents.
It costs more to fabricate LEDs than ICs occupying the same amount of real estate. One reason is that LED chips are made on smaller wafers and, thus, entail higher handling costs. Most high-volume ICs are cut from silicon wafers that are 300 mm in diameter. In contrast, high-brightness LEDs are made on sapphire wafers, not silicon, and the wafers themselves have diameters about half that at most, with many production processes still using 50-mm.-diameter wafers (about two inches). At least one supplier has demonstrated a 300-mm. sapphire substrate, but not as a production item.
LEDs also use a fabrication process unlike that of ordinary ICs. Metalorganic chemical-vapor deposition (MOCVD) is employed for producing thin films of compound semiconductors, as typically used in LEDs, through chemical reactions on the surface of the substrate. The reactions take place at temperatures above about 800°F at pressures up to about 100 kPa in the absence of oxygen.
LED wafer sizes are rising. The market research firm Yole Développement, Lyon, France says more than 65% of the MOCVD reactors shipped in the latter half of last year were set up for 100-mm.-diameter wafers or larger. The firm expects that for the first time in 2011, 50-mm. wafers will represent less than half of all LED substrates. The firm also thinks this could tighten the supply of sapphire wafers because larger wafers are thicker than their 50-mm. counterparts and tend to produce lower yields. (Volumetrically, a 150-mm. sapphire wafer is the equivalent of seventeen 50-mm. wafers.)
Page 2 of 3
Semiconductor manufacturing equipment trends also point to lower LED prices ahead: LED production capacity is growing quickly. Last year, the semiconductor industry organization SEMI recorded a big jump in spending on equipment for LED fabs. SEMI expects that the trend will continue this year with a 40% year-over-year increase in sales of LED fab gear. A lot of these systems are going to China, thanks to the subsidy programs there. SEMI recorded 19 new LED fabs starting up in China last year, with another 27 new fabs expected in 2011. Though high-brightness LEDs get most of the headlines, SEMI expects strong demand from LCD backlighting to drive manufacturing equipment growth for the time being.
Of course, one way to wring out costs is to further automate the manufacturing process. But one difficulty facing LED makers trying to automate is that there are few standards for carrier and equipment interfaces on LED fab equipment. That makes it tough to implement any kind of material-handling automation because each fab is liable to handle wafers with slightly different hardware. Partly to overcome such obstacles, a group called the SEMI North American HB LED Standards Committee formed to help set form factors for widely used manufacturing equipment. Observers say the outcome could be greater use of robotic wafer handling, automated glove boxes, interbay automation, and Standard Mechanical Interfaces (SMIFs). Because LED wafers aren’t as big as the 300-mm wafers used for ICs, the material handling equipment involved needn’t be as heavy.
There is also a move to make LEDs on silicon substrates rather than sapphire bases as a means of leveraging the large installed base of silicon chip manufacturing equipment. For an example of what this development may bring, consider that LED maker Bridgelux plans to slash LED costs by manufacturing gallium-nitride (GaN) LEDs on 200-mm. silicon wafers. To build LEDs on silicon rather than on sapphire, it deposits a special barrier layer between the silicon and the GaN. This scheme would let Bridgelux process LEDs on a lot of under-utilized silicon wafer fab houses around the world. The company sees this gambit as leading to a $5 LED light bulb as soon as 2014.
Nevertheless, LED light bulbs with a single-digit price are still in the future. Until then, LEDs for illumination are seeing use applications where their other benefits cancel out their higher price. For example, a 40-W equivalent LED bulb with a decent hue retails for about $20, and higher-wattages cost even more.
The DoE is well aware of the cost differential of solid-state lighting. At one of its recent SSL workshops, Radcliffe Advisors consultant Fred Welsh demonstrated that commercial 60-W equivalent LED lights have a relative cost of about $50/kilolumen. This is down significantly from even a year ago and is projected to hit $2 to $3 within the next few years, ultimately down to about $1 by 2020. For comparison, the 60-W incandescent lamp that such a light would replace comes in at $0.5/kilolumen. A simple CFL hits about $2/kilolumen. Though LED chips are still expensive, much of the cost reduction is expected to come from more economical packaging, optics, mechanical/thermal components, driver electronics, and a more automated assembly process.
LED lifetimes are improving thanks to better thermal management. One example of this is the development by ZemosLED in Korea of an LED said to last 80,000 hours. (Illumination-grade LED lamps typically advertise lifetimes between 50,000 and 25,000 hours.) The ZemosLED claim to fame is a low thermal resistance between the LED chip (from Cree) and the base on which it mounts, which serves as a heat sink. Most LED packages use a coating of silicon on the LED to maximize the amount of light emitted from the LED and to support the LED bond wire somewhat. Zemos contends that the silicon can become brittle with heat and interfere with heat sinking. Its design mounts the LED chip directly on a copper base which is also larger than the usual copper base. The resulting package keeps the conformal coating cool enough to prevent eventual brittleness over the life of the light.
Other factors besides heat can degrade light efficiency. An example of what happens in most fixtures comes from Cree global director of applications engineering Mark McClear. He points out that cool-white 6,000-K LEDs are readily available with efficacies of 160 lumens/W for use in luminaires. But the LEDs lose some 10% of this efficacy due to thermal issues, another 10% in the lenses and other optical components, and 15% in the driver. So overall, a luminaire designer can expect an efficacy of about 125 lumens/W. Not bad, seeing as LED luminaire efficacy was 50 to 90 lumens/W at maximum drive current six years ago.
Tests performed by the DoE under its CALiPER (Commercially Available Light-Emitting Diode Product Evaluation and Reporting) this past June found that the overall average efficacy levels for solid-state lighting products (LEDs) tested most recently was 46 lm/W. This compares with overall benchmark average efficacy levels of about15 lm/W for 100-W incandescent bulbs tested.
Page 3 of 3
Problem is, the lumen output and efficacies of LED products do not always match laboratory tests of performance. This despite the fact that standard testing protocols have been developed for light output, color, and efficacy. To address these problems, DoE has rolled out a Lighting Facts label to report performance in a standard way.
Thermal stress caused by heat build-up in luminaires is partly to blame for real-world performance falling short of lab results. LED manufacturers have begun to address this problem by providing “hot” performance data on their LEDs. One other problem with luminaire testing is that there is no standard way of making accelerated reliability tests. That’s important when LED lifetimes stretch into the tens of thousands of hours -- After all, the lifetime test of a 50,000-hour luminaire would consume over five years. Such tests, capable of providing accurate projections of life, do not currently exist. So accelerated LED lamp testing remains an area of intense interest
The latest round of CALiPER testing would seem to indicate that luminaire makers still give themselves the benefit of the doubt when it comes to reporting performance. CALiPER Round 13 looked at three types of LED luminaires for commercial and industrial applications: high-bay luminaires, wallpacks, and 2x2-ft troffers.
Commenting on Round 13, the DoE said that accurate reporting and product literature are still concerns - not only for LED lighting products, but also for their benchmark counterparts. Most products were found to have accurate manufacturer claims. But they noted problems with claims of equivalency where LED products were said to be the equivalent of incandescent models of a given wattage.
Accurate manufacturer claims tended to include detailed photometric performance specifications that referenced LM-79, the approved method for electrical and photometric testing of solid-state lighting, and avoided the use of equivalency statements, says DoE. Products that omitted this detailed photometric data and made vague equivalency claims tended to fall short of expectations, the agency says. Only a portion of the products going through the CALiPER tests carried equivalency claims, but most of those claims were found to be misleading or false.
Nevertheless, DoE says several good trends emerged from Round 13. The average luminaire efficacy continues to increase, and color quality continues to improve. Average luminaire efficacy of the LED products exceeded 60 lm/W, says DoE. And while there were still some LED products that didn’t perform well, particularly with regard to light distribution, testers noted much more uniform product performance than in previous rounds. DoE further commented that many products tested took advantage of the inherent strengths of SSL to get a uniform light distribution resembling that of conventional luminaires.
Technical challenges remain
Though LED makers are focused on cost, they haven’t stopped work on surmounting technical challenges in other areas. It is well known that driving an LED with higher current results in more light output. Thus lamp makers would like to drive LEDs with as much current as possible to maximize the amount of LED light. But this scheme has its limits. Higher drive currents reduce efficacy levels because of the well-known phenomenon known as droop. There are still differing opinions about what causes droop, so it remains a major problem area.
LED makers are also putting a lot of work into making devices that are good at producing pure white light. Some kinds of LEDs configured to emit white light lose efficiency because their light must pass through phosphors used to convert LED emissions into the right light wavelengths. Phosphors can also shift color over time and thus change LED light emissions. And even slight variations in the deposition of the phosphors can make a noticeable difference in the color of the light LEDs emit.
Difficulties associated with phosphors have led many LED makers to embrace an alternative means of producing white LED light by combining multiple LEDs emitting light at the right emission wavelengths into a single package. This scheme has its own challenges. It is difficult to maintain consistent white light because of natural variations in the LED blue wavelengths. Some blended LED techniques also use phosphors to down-convert LED light to the right emission wavelength and have phosphor deposition issues of their own. And all white light techniques are susceptible to color shifts caused by variations in LED optical power, peak emission wavelength, and operating temperatures.
Moreover, green and amber emitting LEDs have a low efficiency that currently limits the color-blending approach. To compensate, LED makers may have to resort to special optical designs to effectively mix the light output from the multiple color LEDs. This approach can also require use of complex control circuitry to keep LEDs emitting white light. Further complicating matters is the fact that different-color LEDs respond differently to variations in temperature and current density, and degrade at different rates over their lifetimes. There are ways to allow for these effects through control circuitry, but the result is less energy efficiency.
Also of concern is that LEDs exhibit a slow but significant drop in light output over time, dubbed “lumen depreciation”. However, the character of this depreciation can depend greatly on a variety of factors that include the specific package design, mechanisms intrinsic to the LED chip, the phosphor material, lens material, and other issues related to the assembly of the LED package. Thus it is tough to predict how light from a particular LED will diminish through its useful life. There are standard methods for measuring how light output changes over time, but there is no consensus about how to extrapolate such results to estimate the drop in LED output over the kind of lifetimes now advertised for commercial LED products.
Lumen depreciation isn’t the only LED failure mechanism that concerns manufacturers. There are reports of manufacturing defects in the areas of connection or adhesion failures of LED packages, optical lens failure, inadequate materials control, and moisture penetrating the LED package. And these are just in the LED package itself. When LEDs are incorporated into a luminaire, many additional failure mechanisms may rear their ugly heads. EE&T