0
\$\begingroup\$

This is generally quite a vague question as I'm new to the space but I've been thinking a lot about prototyping in the IC/semiconductor space and have been wondering about the feasibility of developing a small scale/benchtop system that is capable of making cmos chip prototypes.

Can anyone who knows more about this fill in the gaps here and tell me where I'm likely wrong? Why or why not would a system like this work? How would one go about building something like this?

\$\endgroup\$
14
  • 1
    \$\begingroup\$ Have you ever seen how the actual ICs are manufactured? Clean rooms and millions worth equipment? \$\endgroup\$
    – Eugene Sh.
    Jun 15 at 17:06
  • 2
    \$\begingroup\$ Other end of @EugeneSh. scale - garage lithography by Applied Science. Link: youtube.com/watch?v=YAPt_DcWAvw \$\endgroup\$
    – winny
    Jun 15 at 17:08
  • 1
    \$\begingroup\$ @EugeneSh. There actually are desktop scale processing cells for prototype IC production. A whole setup needed to do basic cmos process costs maybe $1.5M and runs in normal office air. The cells are gas tight and the wafers are moved around in gas-tight cartridges. No clean room necessary. At this point, all this tech needs is volume to make it cheaper. The biggest cost long run will be process chemicals and their disposal. It’s surprisingly expensive. \$\endgroup\$
    – Kuba hasn't forgotten Monica
    Jun 15 at 17:17
  • 1
    \$\begingroup\$ @Kubahasn'tforgottenMonica Would you be able to share more details around desktop scale processing cells? Any links? \$\endgroup\$
    – sam
    Jun 15 at 18:56
  • 1
    \$\begingroup\$ @sam The gases (contrary to what Kuba mentioned) are also quite cheap ($50 two decades ago bought a bottle of Arsine from California.) But they are VERY dangerous and it is all the stuff you need to make the environment safe for their application, which will kill your budget. That, plus all the regs related to owning and using them. Silane, Arsine, Phosphine, etc., are EXTREMELY dangerous gases. Used to be the gases were not regulated across state boundaries. I think that's changed. You have lithography and all the processing between steps. In all, it's expensive to set up at home. \$\endgroup\$
    – jonk
    Jun 15 at 19:51

1 Answer 1

Reset to default
4
\$\begingroup\$

Coincidentally, Sam, you want to learn from another Sam, Sam Zeloof.

He's managed to produce one or two devices a year with a few thousand transistors. This is really impressive, but still not all that useful for non-toy purposes.

However, I don't think this is something that can be deskilled and productized like a 3D printer. The chemistry is pretty toxic as well.

\$\endgroup\$
3
  • \$\begingroup\$ "The chemistry is pretty toxic as well." I worked at a defense lab that had a group with their own small-scale fab lab. One guy worked over the weekent, and on Monday was found dead. He had apparently inhaled arsine or phosphine, both used as dopants. \$\endgroup\$
    – WhatRoughBeast
    Jun 15 at 18:58
  • \$\begingroup\$ Out of curiosity - can you elaborate on why deskilling this to something like a 3d printer is infeasible? What are the main constraints around doing something like that? \$\endgroup\$
    – sam
    Jun 15 at 18:58
  • \$\begingroup\$ @sam The success of modern CMOS relies completely on the ability to grow (not deposit, not print) extremely thin and extremely high quality oxide layers. It is simply far beyond any 3D printing technology. \$\endgroup\$
    – Elliot Alderson
    Jun 15 at 20:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.