Visible Light-Emitting Diodes: Past, Present, and Very Bright Future
- PDF / 211,796 Bytes
- 5 Pages / 612 x 792 pts (letter) Page_size
- 14 Downloads / 167 Views
t. The question remained as to whether phosphorus could be added to GaAs to produce an efficient, direct-bandgap material that would emit light in the visible range. Scientists had speculated about this, and most thought that mixing such materials would only create an alloy with so many defects that it would emit little or no light at all. However, Holonyak was able to do it; he generated red LEDs and, in fact, made a red laser.1 In 1962, General Electric sold the first commercial LEDs for several hundred dollars. From that time, the field has evolved tremendously, as shown in Figure 1. Performance is shown in lumens per watt (lm/W, visible flux divided by power into the device), which is the standard unit used
in the lighting business. A 100-W tungsten bulb emits between 1500 and 2000 lm, or 15–20 lm/W. Different materials systems have since been developed, and the rate of improvement in performance has been about a tenfold increase in light output per decade, with the best result being a 100 lm/W research device demonstrated in late 1998. In the last decade, both the AlInGaP system for red and yellow2,3 and the InGaN system for green and blue4 have emerged, enabling the development of very bright devices. Also, some very significant progress has been made in the organic LED field. I do not see these as competing technologies, necessarily. The organic LEDs are driven at current densities three or four orders of magnitude lower to obtain reliable devices. They are suitable for applications such as automotive displays, but not for high-intensity use at this time. I believe the organic LEDs compete primarily with LCD displays. Figure 1 also compares the performance of different types of light sources with LED performance. Note that an unfiltered incandescent bulb emits, as previously mentioned, around 17 lm/W. However, if we put a red filter on it, we lose threefourths of the light and end up with only about 4 lm/W. As soon as red LEDs became brighter than that, we suddenly began to see them in new applications such as automobile brake lights and traffic signals. Some other higher-performance lamps, however—the fluorescent lamp in particular—have performance of 80 lm/W or higher. It will be a challenge to get white LEDs up to that level.
History In 1907, a man named Round took a piece of carborundum, which is silicon carbide, ran a 10-V current across it, and saw all the colors I will be mentioning later. That discovery occurred nearly a century ago; yet, not much further research into this phenomenon was done until 1960. In 1960, Holonyak,1 who was then at General Electric, started researching compoundsemiconductor alloys. He and others had been working with gallium arsenide (GaAs), which is a direct-bandgap semiconductor. It was known that by inserting a p-n junction into GaAs, infrared radiation was emitted. Researchers were also aware of gallium phosphide (GaP). GaP is an indirect-bandgap material that does not emit much light, but because its bandgap is larger, it has the potential to emit visible
MRS BULLETIN/
Data Loading...
