9.1.2 Optoelectronic Devices

Remarks

This module will start in a very superficial way. It will be neither complete in the sense that all important devices are mentioned, nor exhaustive in the sense that the most important specifics of the devices will be included
In time, it is hoped, this module will become more volume and depth from the work of the students in the accompanying seminar.
 
Light Emitting Diodes
   
Light Emitting Diodes or LED's are the cornerstone of optoelectronic products. Very roughly they are used for two main product lines and for one speciality:
1. Signal lights All those little red, yellow, green or blue lights (sometimes blinking annoyingly), mostly used for indicating that something has been turned on, is in a certain mode, or simply is just there (e.g. blinking red bicycle lights).
The main requirement for signal light LED's is that they come in many colors (so designers of dashboards etc. are not limited in their creativity; an important condition considering that designers appear to be limited in many other respects) and that they are cheap.
Nowadays (2008) we do have all colors - including "white" - but that was not always so. For generating a specific "color" including white you can go three routes:
  1. Take a semiconductor with the appropriate bandgap. This generates a "true" color, i.e. light around one wavelength as given by our "master diagram" from before.
  2. Take semiconductors that generates UV or at least high-energy light and use it to excite some fluorescent material - exactly like in fluorescent light tubes. That can produce white light or any color you find a fluorescent material for.
  3. Take three semiconductors that produce "RBG", i.e. red - green - blue, in such an intensity mixture as to produce the color wanted - exactly at it is done by any screen or display.
Obviously the first way is potentially the cheapest as long as you don't require white light. That's why you mostly find LED's belonging to the GaAlAs family (red), GaP family (green) or the nitride family GaAlN (blue to UV). Going from red to blue / UV also mirrors the history of LED development.
Efficiency of LEDs vs. time
Source: M.R. Krames et. al., Applied Physics Letters, 75[16], pp. 2365, (1999)
and Scientific American, Feb. 2001
The orange LED dscribed in the article mentioned above already had an efficiency of > 100 lm/W, and today (Feb. 2008) 150 lm/w white LED's can be purchased, for example from Nichia (the company that pioneered the blue / white LED). But now we moved already into the second topic:
2. Light With the advent of the blue / UV GaN-based LED in 1993 (that comes with a quite interesting story around its inventor, Shuji Nakamura and the company he was then working for (Nichia)), making white light with LED's was possible for the first time.
For general lighting purposes - your room, the street, a congress hall, the street in front of you bicycle or car, - you name it - you need first of all white light. After you can do that with LED's, you need:
  • High efficiency as measured by lumen / Watt (lm/W) or by total "plug efficiency" in %, meaning the ratio of light energy out to electrical energy UIt in.
  • High intensity. It is not good enough to have a high-efficiency light source if the best you can offer produces the same intensity as, let's say, a 20 W conventional light bulb.
  • Large life time. You do not want to change your light fixtures too often. For something in the better quality region, you want several years of operation time at the least.
  • Low price. The price you can get for your white light LED depends on what you offer. If it is much better than a regular light bulb, it does not have to be "cheap" - but it still must be worth its price.
The picture below shows the efficiency of white LED's vs. existing light sources:
Development of light source efficiency
Based on information provided by Osram
History of Light sources incl. LED
The insets symbolize the type of "light bulb" sufficiently; the one belonging to the purple line (the top performer of the "classical" light sources) is the metal halide "bulb", belonging the the "high intensity discharge" (HID) type of light source. If you compare the development of the white LED to all the other light sources, you get a first impression why everybody in the lighting business is so excited about LED's as the light source of the future. Potential energy savings are enormous!
A quick word to the unit "lumen" and to absolute efficiencies:
The "lumen" (lm) is the SI unit of the perceived power of light; it measures the luminous flux. Now the natural way to measure the flux of light would be to measure the energy flux of the light. Since the eye, however, is not equally sensitive to the various wave-lengths it can see, the lumen corrects for this.
As far as converting lm/W into absolute efficiency goes, we have the following approximate relations
  • (10 - 15) lm/ W Û (5 - 9) %
  • (70 - 100) lm/ W Û (25 - 35) %
With respect to efficiencies, LED's do have a bright future, indeed. What about the other criteria?
Product life time is not a problem but an asset. Well-made LED's will last for >10 years, outperforming more or less all other light sources.
The problem is: intensity. The light is typically produced in a small volume (great for focussing etc.), and if you put a power of let's say 100 W into a volume of <1 mm3, you better have some concept of keeping the temperature down.
Closely related is the problem of plug compatibility, meaning that you want to use 230 V 50 Hz AC, 110 V 60 Hz AC, or whatever your country has as its consumer electrical energy standard. A LED, however, is a forwardly biased pn-junction, running at something like 3 V DC (and then 33 A if you want 100 W). While this is an electrical engineering problem, it is still a big problem.
The (inorganic) materials used for both applications (always as thin layers) are once again:
  • Aluminium gallium arsenide (AlGaAs) — red and infrared
  • Aluminium gallium indium phosphide (AlGaInP) — high-brightness orange-red, orange, yellow, and green
  • Gallium phosphide (GaP) and Aluminium gallium phosphide (AlGaP) — green
  • Gallium arsenide phosphide (GaAsP) — red, orange-red, orange, and yellow
  • Gallium nitride (GaN) and Indium gallium nitride (InGaN) — near ultraviolet, bluish-green and blue
   
Laser Diodes
   
Semiconductor Lasers will be treated in some more detail in module 9.2.2. Here we simply note that the theory of (semiconductor) Lasers is rather complex, but the technology is not.
   
In principle, many LED's "automatically" become a Laser if you run a very high current through them (producing a lot of light) without destroying them first. It is therefore not too difficult to produce a Laser diode in the Lab - all you need (haha) is very efficient cooling of your experimental LED device.
CD Laser head
Laser and optics of CD player.
You may know already that Lasers always need some kind of optical feed back, usually provided for by mirrors, and ask yourself where the mirrors are if we use a simple LED as Laser. The answer is that plan-parallel surfaces of the semiconductor might be already sufficient for that because the interface semiconductor - air does act as a "semi"-transparent mirror and that might be good enough.
The truth, however, is that it often takes many years after a certain new LED has been marketed, before the long-lived, reliable and cheap Laser diode follows.
The first GaN-based blue LED's were on the market around 1993, whereas the blue GaN based Laser had to await 2005 or so (in 2006 SONY, for example, still had major production problems).
Using semiconductors for making a Laser is just one way for making Lasers - you can use other solids, liquids and gases for that. This brings up the question of pro and cons - what are the advantages and disadvantages of semiconductor Lasers?
Look at CD and DVD or now blue ray disc drives - the are the major market for semiconductor Lasers besides the very pedestrian "Laser pointer". The advantages are obvious:
  • Very cheap.
  • Very small.
  • Electric energy supply at low voltage.
The major disadvantages are
  • Low power at decent quality (around 1 W maximum).
  • Limitations as to color (= frequency).
The cheap and reliable semiconductor Lasers are actually the enabling devices for all this memories! No suitable Laser - no discs.
Sorry - your Laser pointer just can't be turned into the Laser gun you might fancy for fighting those aliens if you are a male (or, in your adult life, for cutting metal or other materials). For this you need other Laser types which can deliver real power - far heavier, bulkier and far more expensive than your semiconductor Laser.
Of course there are constant optimizations and new developments - newer and better semiconductor Lasers are frequently announced. We haven't seen the last of this yet.
Displays and OLED's
   
Take a million or so LED's, arrange them in a matrix, and make sure they can be individually addressed - you now have a display.
   
If you make your display with individual LED's, soldered together somehow, you will get an expensive big display with lousy resolution - the kind of boards you see on Times Square or other places that have been taken over by the evil advertising people.
OLED flat panel display from Samsung
OLED based TV
Courtesy of Samsung
If you could make your LED's tiny in the lateral direction and all of them close together, i.e. on one substrate, and for all three RGB colors and individually addressable, you would have a flat panel display that would be great for TV, computers, cell phones and laptops because it could have a high energy efficiency.
Unfortunately, the possible substrates for inorganic semiconductor LED's are far too small (we have only the III-V single crystals, essentially GaAs, SiC and perhaps Al2O3 (= Sapphire)) and those potentiial substrates do not come even close to what would be required.
Fortunately, an unexpected discovery made accidentially in the 70ties by Shirakawa in Japan (and to some extent by others before him) has helped:
There is such a thing as an organic conductor, and an organic semiconductor leading to an organic light emitting diode - an OLED.
OLED's have only be around for less then 5 years, but we already have flat panel displays based on OLED's in cell phones and the first ones for TV are announced right now.
Let's be clear about one thing: Organic semiconductors right now are still lousy semiconductors (and extremely sensitive to oxygen).
They have a tremendous advantage over inorganic semiconductors, however: they can be very cheap and, far more important, they can be deposited at low temperatures by rather simple techniques on cheap and very large substrates and they are easy to pattern.
In other words: OLED displays are easy to make, the light emission is OK, and the product life time is OK at present for consumer items where demands are a bit more relaxed.
What we are witnessing right now is the beginning of a completely new field of semiconductor materials science and technology. Who knows where it will end!

With frame Back Forward as PDF

© H. Föll (Semiconductor Technology - Script)