LED Fabrication (Take 2) & Vacuum Technology

So when I previously wrote about LED fabrication, I talked about doping. It is, in theory, one way to create the desired material properties to turn electrical current into light. However, the efficiency of an LED depends on how crystalline (how orderly or consistent) the device is. And doping is like…..well….shooting bullets into a nice bale of hay. It’s not very crystalline. It’s a great way to make electronic gates when you don’t have to worry about high current flow, because you are more worried about pushing a force verses an electron through the device.

So instead of shooting atoms into a material, we would rather coat them in thin layers. And this is typically done by a process called MOVPE (or metal-organic vapor phase epitaxy), or just VPE for short. They can also use MOCVD (metal-organic chemical vapor deposition), which is kind of similar. And there are many other theoretical and experimental ways to layer atoms, but there are only a few, well, feasible of creating an LED for 5 cents.

So this is the basics behind MOVPE. Taking a material that usually wants to stay a solid (the metal “stuff”), we combine it with an additional molecule (an organic compound) which makes the MO material more volatile (basically a vapor phase). A lot of times it’s ammonium (NH4+, but anything goes as long as it works.

The vapor coats just about everything in a thin layer of the metal-organic substance (wafer, pipes, chamber, you name it). And the fancy word for atom-layering is “epitaxy.” So there’s the five parts to that massive scientific acronym. Change the vapor pressure, temperature, flow rate; and the growth rates are altered in non-linear, but predictable ways.

You only need 2 layers to create some sort of LED, but that’s not exciting enough. For, example, modern chips create a mirror underneath the LED. And when I mean “underneath, they grow the LED, sputter metal/oxide/something on top, and flip the chip around. You can’t grow anything “underneath” a crystal layer without breaking it apart, so there are a lot of fancy fab tricks that we utilize.

The whole topic of high-vacuum processing is interesting by itself. When the only atoms in the chamber are the ones of interest (metal/oxides) without “air” in the way, they behave very differently than what we are used to. It’s like the difference between working with water and sand (if that makes sense) They don’t interact with much besides the vacuum chamber itself and the most important part, what we put in it that vacuum chamber.

We even have to create vacuum pumps in a very different way to reach sub-atmosphere vacuum pressures, because we aren’t actually “sucking” at that point. Hyper fast turbines with hundreds of fragile blades, ones which would rip themselves to shreds in normal pressures, force single atoms through numerous rebounds. Cryogenic pumps cool down atoms to bring them out of the vapor phase. Non-volatile oils, which don’t evaporate even under extremely low pressures, “flow” in circular paths to catch unsuspecting gas molecules.

And once you have removed all the unwanted “junk” from your system, then you can have some real fun. Wondrous things can occur when you don’t have “gas” in the way. Cathode ray tube technology still exists in many forms. Nothing like glitter when you bombard an electron layer into a sample of pure gold. Or utilize it to raster scan a sample to create your ever-so-popular “Scanning Electron Microscope” (SEM) images.

Or you can add some pure gas, and excite it with some electromagnetic radiation. You don’t know fab unless you have seen the eerie purple glow of Argon as it eats away at your target sample. And specific chemicals will selectively eat specific materials. Oxygen eats plastic; chloroform eats specific semiconductor compounds; and sulfur hexafluoride (which is a lot of fluorine)……… yeah, I’ve seen that used too. Or if you take SEM technology, except you replace the electrons with actual atoms (Gallium, for example), you can physically mill into your sample. Just like this 10 micron flower that I did last year.


Is there a practical reason why I made the flower….not really. But it made for a very unique Mother’s Day gift, even if she doesn’t truly understand how I made it.

But it’s one of the few ways to truly create sub-micron structures out of nothing. You can’t just take the tiniest drill bit and “mill away.” Yes it’s slow, but by making a “master,” you can copy the sample into “daughter” wafers significantly faster. Stamping family seals into wax is significantly easier that making the wooden/rubber stamp itself. But you still need to make the stamp. That’s nano-imprint lithography for you in a nutshell.

How do I wrap this up……

While reading and studying the theory behind modern-day applicable science is interesting (while also putting you to sleep), it doesn’t even come close to physically experiencing these ideas put into practice. Even the “mistakes” are memorable, like my partially reflective mirror of a glass slide (not enough metal), or my wasted efforts in vacuum-based “Reactive Ion Etching” (lots of glowing gas). All while watching my advisor’s money going down the drain….kind of. But you don’t know unless you try, and that’s why we call it research!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s