1990-1999: LEDS BREAK THE QUANTUM BARRIER
This is a close-up of the die (light-emitting region) of a ZnSe blue LED.
Photograph was used with permission of the person who furnished it.
I don't know much about blue ZnSe LEDs and I've never actually handled or even seen one; that's why you don't see a big long text about them here.
PDF file about ZnSe LEDs
File is 1.1MB (1,243,467 bytes) in length; dial-up users please be aware.
File has also been used with permission of the person who furnished it.
(Sent by a website fan, received on 05-16-05)
This is a blue-green LED with an approximate date of manufacture in the mid- to late-1990s. It appears to be a Nichia NSPE590S.
Measures 37,000mcd on a Meterman LM631 light meter.
Measurement appears high because the LED has a very narrow beam.
Vf is 3.58 volts at 19.28mA.
Here is a photograph of its beam on a ceiling from ~5 feet. Deliberately underexposed by two stops to show the beam.
That small yellow square thing under the 5mm LED is a 1990s-era yellow-green LED. The person sent me several; only one is shown here. They came out of a Sheikosha EDM Survey Machine, from about 1997. He wanted the optics, about 30 X, and the waterlogged EDM was to be binned (thrown in the garbage), so he dismantled it and rescued the LEDs.
Since Henry Round's discovery in 1907, and Russia's attempt in 1973, little has been done with silicon carbide other than to coat sandpaper and filing wheels with it. That all changed just after 1990, when the first stable specimens of blue LEDs came to life.
You see, until the blue LED, there was no way for humans to create full color displays, message boards, or other full color LED lighting.
With this technological breakthrough, it was now possible to make all these things, and more. These early blue LEDs even led to the creation of the world's first white LED - which comprised of two SiC blue chips, a GaP green chip, and a GaAlP red chip all stuffed into a single, standard sized LED package. An example of that LED will appear below.
This particular blue SiC LED has a light output of approximately 5mcd, which is very weak by today's standards.
But it was enough - the blue LED was born. I was one who partook of this new development, and had the first computer in town with a blue power-on light. :)
(Sent by a website fan, received on 05-16-05)
This is a blue SiC LED in a diffused, 5mm epoxy case. It has a light output of approximately 5mcd, which is very weak by today's standards. It has a Vf of 3.09 volts at an If of 19.28mA.
Awwww, the poor thing... :(
This poor fella is also a Silicon Carbide blue LED, but it's not exactly in pristine condition anymore.
Purchased for $15 in late 1991 or early 1992, it was in a piece of equipment which was struck by lightning some years later.
And even though it's blackened and burnt inside, it still has a glimmer of life left in it.
The left picture is of the LED as it is today, appearing exactly as it was pulled from the destroyed computer.
The center picture is a look right down the barrel. Most of the insides are shades of black and brown, but some turquoise blue
light still manages to escape in places.
And on the right, the LED's glow as you would see it looking from the side.
When new, this LED was rated for 50mA of drive current, and had an intensity of 14mcd. It was actually a rather pretty shade of blue;
a cross between azure and sky blue. It is (or shall I say, was) the brightest SiC blue I've ever come across.
This LED was the first true RGB LED to hit the market. In addition to being able to produce hundreds of different colors, if you adjusted the current to each color just so, a reasonable, albiet weak, approximation of white could be created.
The milky, diffused case helped mix the individual colors, and was reasonably effective either at longer distances, or when the LED light was bounced off something white. Brightness appears to be around 10mcd with full power going to both blue chips, and the red & green adjusted to give an overall white color. The red and green chips were much brighter than both blues combined, so you really had to design well to get good color rendition out of this LED.
A modern RGB LED, using the brightest red AlInGaP chip and InGaN green & blue chips easily outpowers this LED by a hundred times or more, and is the current standard for use in smaller sized video boards, scrolling message signs, and other smaller to medium sized full color displays. Physically, the modern LED resembles this older model, but has just four leads (fewer chips to drive inside), and not six.
This is arguably the first Indium Gallium Nitride LED to show up outside a laboratory setting. It is a LEDTronics model L200-CWGB6-100. Look closely...
what do you see? More precisely, what don't you see?
That's right! No bonding wires at all. The very large crystal of either silicon carbide or artificial sapphire doped up with a gallium nitride emitting layer was simply
flipped upside down and clamped onto the surface of the LED's metal leadframe!!
The faint, silicon carbide-like glow came from a small region on the cathode where contact was made between the metal leadframe component
and the GaN slab. This LED's peak wavlength was approximately 485nm, which is a turquoise shade of blue; not that unlike most of the earlier silicon carbide LEDs.
It was originally intended to be used in a colorimeter. Soldering to this LED would have been extremely difficult and result in a high mortality rate, as every single microdyne of
stress placed on the LED's leads was transmitted directly to the poor chip inside. Heat from soldering would soften the plastic of the LED case, only compounding the problem.
The only way this LED could have survived installation in a commercial application is by socketing, and even then, a strong enough hit to the cabinet of a machine equipped with
one of these (or a fall from the lab bench to the floor) just might be enough to kill the poor LED.
According to my source, this LED was made in 1993 or 1994 and sold for only a short time before brighter and more robust GaN LEDs began to come out of Shuji Nakumura's lab at Nichia.
Check out that pebbly texture in the region emitting the light in this magnified top view. This is probably indicative of a clamp of some kind that applied tremendous pressure
to the GaN chip in order to make a bond with the metal leadframe structure it is mounted to.
The spectrum of this LED was unusually broad. At lower currents, this and even most modern GaN and InGaN type LEDs tend to have very broad spectrums with a lot of green & red and not much blue.
At its nominal operating current of 10 milliamps, the color would have shifted noticeably toward the blue and the emission in the red would not increase nearly as much in relation
to the rest of the LED's output.
More importantly, this picture shows how large the LED die is in relation to the rest of the the structure, and the small size of the actual emitting region compared to the size of the
die as a whole. Even by today's standards, this would be considered a large junction device, even though only a small portion of it is emitting any energy.
The other remarkable feature is that this LED had a Vf of 11.0 volts (!!!) at 10 milliamps.
Photos of this LED courtesy of Paul Schick.
As of 05-06-03, I now HAVE one of these wily and elusive LEDs, and I didn't have to bash open an old colorimeter to get it!!!
Stay tuned to this station for additional pictures!!!
Spectrographic analysis of this LED at an If=10mA.
Spectrographic analysis of this LED again; spectrometer's response narrowed to a range of 420nm to 650nm.
USB2000 spectrometer graciously donated by P.L.
Here is an early Indium Gallium Nitride LED. First appearing in 1996, this was the first blue LED that could be classified as being "ultra bright", and this was the LED that led to today's single-chip white LEDs.
People who came across one of these got another surprise when they looked inside: instead of a tiny cubical chip, this LED has a large, flat square chip nearly 1/4mm on each side. The other distinguishing characteristic that you can see is how the chip gets its power: instead of a single thin wire attaching to a tiny gold ball in the center, this LED's chip has two wires connected at the corners of the chip. This was done because the substrate (the material the LED "stuff" is built upon) is artificial sapphire, and it does not conduct electricity. So wires had to be run for both the anode and cathode and connected directly to the top of the chip.
In reality, the process is much more complex - if you're interested, look up terms like "MOCVD" or "epitaxy" on a good search engine sometime. This museum is primarily for your enjoyment and maybe some basic education - not your next class in quantum mechanics. :)
And this is the modern-day counterpart to the SiC/GaAsP/GaP full color LED of the early 1990s.
This powerful LED, manufactured by Nichia America, uses the latest InGaN, GaN, and AlInGaP II chips available.
For those new to this whole thing, InGaN is indium gallium nitride and is used for the bright, pure green; GaN is gallium nitride and makes the bright blue;
and AlInGaP II is aluminum indium gallium phosphide which creates the brightest reds available.
For the green & blue LEDs, less indium gives a bluer color. Blue GaN LEDs still have *some* indium, but are commonly known as simply GaN or gallium nitride. Green has much more indium,
and are almost always known as InGaN or indium gallium nitride.
Put all of these superbright dice (light emitting chips) together in one LED body, and you have the world's brightest true RGB LED.
This particular part started showing up in the late 1990s.
Update 12-06-09
This web page shows the "inner workings" of a self-flashing RGB LED; I'd rather highly recommend a visit here to learn a bit more about these critters.
This is the actual LED used.
What's this?
Oh, it's just one way to show just how bright today's blue LEDs really are.
In this picture, they're being used as taillights in an electric wheelchair.
These LEDs were purchased in late 1997, so they are not nearly as bright as those you can buy today.
But... they are brighter than the early MQW LED shown just above. These were the next technological step ahead, which increased blue LED brightness by another ten times.
In the mid 1990s, starting around middle or late 1996, white LEDs started showing up. A white LED is actually a blue LED in sheep's clothing; it uses a blue LED chip and a phosphor that converts some of the blue to white.
The photograph directly above is what a white LED actually looks like under 100x magnification!
This is an example of a late 1990s single-chip white LED.
These are made by taking a blue GaN type LED and covering the chip with a phosphor (a material that glows). What happens
then is the blue light excites the phosphor, which absorbs some of it and converts it into a yellowish, broadband emission that when coupled
with the remaining blue, gives a good approximation of white.
That yellow $@&%* in there - that's the phosphor as it appears when the LED is unlit.
So, have you figured it out yet?
Is this a germanium diode or is it an LED?
Remember, not every item in this game will be an LED.
Looks can be... deceiving!!
???
NOT THIS TIME!!
It's an LED! A very strange LED.
Spectrographic analysis of the Panasonic LN3 LED.
Same as above; newer spectrometer software & settings used.
Panasonic made these bizarre parts at one time. They have a leakage current in the femtoamps and were used as a low leakage reference diode.
Nowadays, they use something called a "bandgap reference" for the really low voltages; above about 1.8 volts we start to see ordinary zener diodes.
The one pictured here is their LN3 green version. I also have the LN2 GaP red version, which you'll see directly below.
Spectrographic analysis of Panasonic's LN2 red version.
The instrument's response was lengthened slightly to 800nm to show the NIR emission.
Spectrographically, the LED formulation appears to be GaP (gallium phosphide) {note the significant "hump" in the yellow-green and the unusually broadband emission in the red extending to the NIR}, just like I said it would be.
Spectrographic analysis of Panasonic's LN2 red version; newer spectrometer software setting used.
The instrument's response was lengthened to 850nm to show the NIR emission.
Thanks for playing. :)
* 1960 * 1970 * 1980 * 1990 * 2000 *
RETURN TO MUSEUM LOBBY
This page is a frame from a website.
If you arrived on this page through an outside link,you can get the "full meal deal" by clicking here.