The way how LEDs generate light means they have no principal energy losses in transforming electrical energy into light. The conversion is really direct (electrons and holes gaining energy by the electrical field in the junction directly turn into a photon once they recombine, no intermediate steps with excitated atoms or so), so less chance to get the energy turned into heat instead of the photon. That is significant difference from all other light sources. However the present design of the LEDs means there are relatively huge losses not related to the main physical principle, it is about 30..70% (mainly because the generated light is getting absorbed by some obstruction structure, like the main body of the chip around the junction, the substrate, metallization layers distributing the current overthe chip area,...; incandescents are ironically the most efficient if only this type of losses would be accounted, as there is no obstruction between the radiating filament). Other type of losses comes from the color conversion (by converting a blue color photon into a red one you loose the difference in their energies, usually in the form of heat generated by the conversion phosphor, in some special cases it is released as an IR. But compare to e.g. fluorescents, these losses are rather low, just because the energy difference between the exciting and radiated photons is not that high like in e.g. fluorescents (~250nm to 500..750nm on fluorescents vs ~500 to 550..750 in LEDs). Still even with these losses the overall efficacy tends to be very high, because most other sources have their principal losses way higher (incandescent emits only few percent of the power in visible, the rest is IR, fluorescents have very high color conversion losses because of the large energy difference between exciting UV and radiated visible),...
That means the classical sources are practically on their maximum attainable eficacy, but LEDs have still very significant room to improve. The principal losses cannot be eliminated at all, the conversion just by not doing that much conversion (using red LED chips instead of a phosphor converting the blue to red part), but that still has to be finetuned (the red LEDs tend to have very temperature dependent output, which kills the clor balance,...). The nonprincipal losses (part of the light source blocks the generated light) are those which could be eliminated by further development. Because these are the only major losses in LEDs, there is expectation the further development will really cut them off significantly. And it is this, where the major changes in LEDs happens now and will for sure hapen in the future, from the present ~60% losses starting point there is huge room for improvement. Once eliminated, the LEDs have the hypothetical potential to be 90%+ efficient electric powered light source, there is no other known light generating method with such potential.
However the LED face mainly the problems of cost and reliability. They need rather complex ballasts. Larger chips are less efficient, so you need a lot of small dies to get the power yet maintain the efficiency. That means a lot of connections and so a lot of potentially failing spots. And a manufacturing cost problem as well - larger source needs just two connection points, so not that much deal if each cost half a cent. But with a LED assembly requiring 100 chips for the same output half cent per connection means you have spend a dollar just to interconnect them together. So you need another, way cheaper method. But at the same time they need to have 100x lower failure rate. The cost and reliability aspects do not like each other, so a lot of development work to be done there as well, hence the other development area for quite a long time into the future.
So yes, it may seems "just a diode", but there is still huge room for improvements to better serve as a light source.
|