[sdiy] Having a hard time finding the chips you want - roll your own!

Scott Bernardi sbernardi at home.net
Wed Aug 29 06:16:43 CEST 2001


Building even a small scale wafer fab unit is not something you can do in
your garage. There are some NASTY chemicals involved - I doubt if you could
purchase them. Boron, phosphine
Here's a list of equipment needed (at least what I can remember - it was
quite a while ago)
Epitaxial ovens
Diffusion ovens (up to 1500 deg C, controlled ramp up and ramp down
temperatures)
Mask alignment and exposure station
Silox etching
Gas delivery and recovery systems (gotta put the doping gas into and out of
the ovens)
Aluminum sputtering units (for metal masks)
Wafer saws

A few million dollars new, maybe a few hundred grand used?

Here's the process for bipolar:
You start with a p-type wafer, then grow a 25 micron n-type epitaxial layer
on top of it.
Grow a layer of silox (silicon dioxide) on top of this
Apply photoresist, expose the isolation mask, develop, and etch the exposed
silox away (just like a PC board!), remove photoresist.
Stick it in a diffusion oven and blow boron gas at it at about 1100 degC to
diffuse p-type material all the way through the epi layer to the p-type
wafer. This defines tubs of n-type material.
Grow another silox layer, apply photoresist, and expose the base layer
(will become base of npn transistors), etch, and strip photoresist.
Another p-type diffusion step to define the bases of npn transistors, and
resistors - islands of p-type material inside of the n-type tubs.
Another silox layer and mask exposure to define the emitter areas.
Another diffusion step, this time with n-type gas (phosphine)
Another silox layer and exposure and etch to define contact openings down
to the components.
Sputter aluminum over the whole thing.
Photoresist and exposure for the metal mask, etch.
Another silox layer, then a mask to define the pad openings (where the
little wires will bond to), etc.

And that's a relatively "simple" process - no ion implantation, etc.



Darren Reid wrote:

> "Magnus Danielson" <cfmd at swipnet.se> sez:
>
> > Hmm... interesting twist actually! But the reason for my questions
> > where to try to shake out some knowledge about what stuff is actually
> > needed. Since I was thinking of pretty large features (by todays
> > measures) leftover equipment should be available for feasable
> > prices.. if one would dare that is.
>
> Shake that tree...I wanna learn more about this too.
>
> -Darren
>
> ---
> Outgoing mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.265 / Virus Database: 137 - Release Date: 7/18/01

--
Scott Bernardi
sbernardi at home.net





More information about the Synth-diy mailing list