7. Light Emitting Devices

7.1 Basic Requirements and Design Principles

7.1.1 Products, Market, Materials, and Technologies

General Considerations

	All light emitting devices – LEDs or laser diodes – have some device principles and requirements in common:
		There is a defined volume in the device – the recombination zone (or active volume) – where the generation of light takes place.
		The increase in the minority carrier density necessary for radiative recombination is obtained by injecting electron and hole currents across suitable junctions.
		The device is made so that most of the injected carriers recombine radiatively in the active volume – i.e. the quantum efficiency and the current efficiency should be as large as possible. This is exactly the opposite of the regular Si p–n junction, where we try to keep recombination in the SCR (and thus leakage currents ) as low as possible.
		For lasers, an optical feedback mechanism is added (e.g. a Fabry–Perot resonator). In addition, the geometric shape is important: Should the laser emit along a line, or just from a "point"?
		The optical efficiency must be optimized, too
		And, not to forget, an important consideration neglected so far: For many applications the modulation frequency range should be large. In other words, we want to modulate the light intensity by modulating the injection currents at high frequencies – GHz, if possible.
	This is a demanding list of specifications; it cannot be met with just a few basic device architectures.
		Considering the wide range of available semiconductors and the extremely diversified product spectrum, there is a bewildering multitude of devices from many materials involving often ternary and quaternary semiconductors from the III-V zinc-blende lattice set.
		While the most complicated devices concern laser diodes (always with spectacular physics involving all kinds of quantum well structures and tricky resonators), the humble LED is not to be sneered at either. It is always the base of laser: If you cannot make an LED for a certain wavelength, you sure like hell will also not be able to make a laser.
		The field was revolutionized some years ago when Shuji Nakamura , almost single-handedly, made working blue LEDs based on GaN; a feat that seemed to be impossible since all the big players in the field could not do it.
		At present, a race for a 12 · 10⁹ $/year market is gaining in speed (and expenditures for research): Cheap LEDs suitable to replace light bulbs, emitting white light at high intensity may be around the corner! But maybe they are not. Only time (after considerable research and development) will tell.

For light emitting diodes (and of, course, for laser diodes even more so), efficiency is of supreme importance: How much light can you get for 1 W of electrical power that goes in the device?

Much progress has been made, and more will have to be made for light bulb replacement. The figure below illustrates that.

The record holder, the inverted pyramid LED, is described in some detail in the link.

How large is the market for III-V devices?

Here is a link to some information.