Friday, May 2, 2008

Notes from the semiconductor lab we visited last week [April 25th]

We visited a semiconductor facility a week ago; this is on the RIT campus as part of the engineering program's apparatus (specifically, Microelectronic Engineering)

The most obvious thing is the 'clean' in 'clean room', and some of the cleanliness measures employed were rather interesting.

It's a Class 1000 cleanroom, meaning them aim for less than 1000 particles per cubic foot.

For instance, outside laptops weren't allowed in (because of dust & related crud that may be in the keyboard), so this post is actually a rough transcription of my paper notes [plus some additional commentary on the lab].
For really hardcore microelectronics engineers, there's special paper that 'sheds' less particulates into the air.
The full-body suits were interesting and amusing-looking, we collectively looked like future astronauts or something.

Air pressure is slightly higher inside, so that clean air rushes out and dirty air doesn't rush in when access doors are opened

The day before our visit was "Bring your Son/Daughter To Work Day", and our tour guide remarked that the gaggle of 5-year-old present looked rather cute after bieng outfitte din the suits.

Sodium is a really bad contaminant. [editor's note: so no salty munchies in the lab!]

------------
Anyway, their equipment is sufficient to produce microchips from beginning to end, but it's all atleast a few years old. (and thanks to Moore's Law, that makes the chips quite out of date; but the value is there as an educational tool.) Furthermore, most of the equipment was donated by
big names in the field (such as IBM and Intel) after *they* upgraded.

Furthermore, consumer-photography companies like Canon, Nikon, and Kodak are also involved in this type of industrial equipment. [The photoresist process, important to semiconductor manufacture, does involve exposing the silicon wafer to a certain amount and type of light in certain areas of the chip.]

Makes sense in a way, because as is common for RIT, grads often end up going right work with one of the big-name companies in their field
-------

So, roughly speaking, how does the process actually work?

* A block of processed silicon is 'sliced' into wafers, and each of those wafers will become a large quantity of chips over the course of the manufacturing process. Thus, each batch of microchips necessarily contains a large number of individual chips.
* With lithography, a printing process, computer-designed patterns are put onto the wafer
* Then we apply photoresist; we apply it over certain parts of the wafer, so that some areas of the chip 'resist' when exposed to light and some won't [this forms the pattern]
* The photoresist is exposed by a certain wavelength of blue light from one of the lab's machines

To prevent accidental exposure, the lightbulbs in that area are yellow and the windows are tinted. I recognized this as being analogous to the use of red-colored lightbulbs in traditional photofilm darkrooms.

* Ion bombardment, towards the end of the process implants material into the wafer that makes it more conductive. (Yes, we discussed silicon for quite a while before mentioning implants. :))
------
Dangerous chemicals are important to the semiconductor process, but one sign in the lab touched upon this with a note of humor:
"You can walk on a wooden leg, you can eat with false teeth, but you can't see with a glass eye. So wear your safety glasses."

No comments: