|
|
Back to UserFriendly Strip Comments Index
|
Moore's Law... just got extended ;-) | by DaNutBall | 2006-11-19 12:55:59 |
| Hmm |
by beez |
2002-06-20 08:58:07 |
Perhaps I'm misunderstanding, but it would seem that this has some drawbacks.
All right, here's my disclaimer: I work for a semiconductor manufacturer, but don't actually do anything directly related to the fab. That being said, I believe I know what I'm talking about here, having spent time in RPI's fab and being a Computer and Systems Engineer/Computer Science dual major.
Problem 1: This process doesn't allow for multiple layers.
Current UV-lithography (and other) processes can have 15 or 20 masks to lay patterns to be etched into the silicon. I'll refer to the type of uv-sensitive material as "photoresist" because that's the brand/type I've used. The exposed areas dissolve when placed in acid (or, depending on the type of photoresist, the non-exposed areas dissolve), exposing the silicon which can then be doped or further etched, etc. Then by dissolving the rest of the photoresist and building an insulative layer (usually SiO2), you create a buffer. A new layer of photoresist can be applied to this layer, and can be etched similarly, leaving holes down to the first layer. This allows for non-planar circuit design, which is quite important in advanced circuits. The process mentioned in above link is (as far as I can tell) not capible of this. It's hard to tell, since the article is very sparce on details.
Problem 2: Uncontrolled SiO2 Growth
When you have a silicon wafer in a oxygen-rich environment, it "rusts", forming SiO2. This is sped up signifigantly by heating. If you plan on heating silicon to it's melting point (1410° C or 2570° F), signifigant oxidation will occur. Unless this pressing process is done in a oxygen-free environment (perhaps in an inert gas like helium or neon), oxidation will occur in the gaps of the quartz pattern and at the edges. I wonder if there is a control process for this.
Problem 3: Head expansion
I don't have time to fully flesh this idea out, but here's the gist: when you heat things, they expand. Of course when they cool, they contract. This has potential to change the shape of the mold and silicon as it heats, thereby causing distortion. You might be able to compensate for this. I don't know, and don't have time to see what the heat expansion numbers for both silicon and quartz are.
Ok, I'm done.
beez |
|
[ Reply ] |
|
You know, I think that's my longest post ever (n/t) | by beez | 2002-06-20 09:47:30 |
|
|
[Todays Cartoon Discussion]
[News Index]
|
|
