4 ms·

I used to work in a fab. As far as I know, there's no acceptable way to remove native oxide without HF. We do have another etch process -- dry etch -- that us

I used to work in a fab. As far as I know, there's no acceptable way to remove native oxide without HF. We do have another etch process -- dry etch -- that uses plasma instead, but it causes surface damage, and you don't want that under the gate. Also, dry etch chambers are a lot more work to maintain than a fume hood where you do wet etch. Furthermore, dry etch has its own set of hazards, and you have to deal with a lot more potentially toxic vapor.

sevensor 9y ago
Thinking a little more outside the box, you might look into using germanium instead of silicon. I don't know much about the germanium system, except that it supposedly grows less oxide. You're still going to have to deposit and remove oxide somehow, though, which means CVD reactors and silane. It's not going to be easy to get away from explosive toxic gases and dirty processes.
- kurthr 9y ago
  Not to mention DiBorane and Arsene... 1000C anneal cycles, teratogenic photoresist ethers, UHV equipment, and muti-kW RF sources. It's possible to demonstrate um scale MOS with a small lab, but not safely in anything resembling a normal garage. Hoping to replicate a multi-$B process at home is a fool's errand.
  - sevensor 9y ago
    While we're at it, a modern scanner costs many million dollars, weighs tons, and sits on its own separate foundation on pilings that go down to bedrock so that it doesn't pick up vibrations from the air handling systems. DIY it ain't :)
    - DoctorOetker 9y ago
      what is a scanner in this context? what is its role? or is this a stepper? or scanning electron microscope?
      - sevensor 9y ago
        A scanner is a photolithography tool like a stepper, except instead of exposing the entire shot at once, it sweeps (scans) a narrow rectangular exposure across the shot, and then steps to the next one. This allows for lighter, more precise optics. Confusingly, there's a piece of standard office equipment with the same name. Even more confusingly, Canon makes both kinds! Be sure you call the right technician for service.

sevensor 9y ago

Thinking a little more outside the box, you might look into using germanium instead of silicon. I don't know much about the germanium system, except that it supposedly grows less oxide. You're still going to have to deposit and remove oxide somehow, though, which means CVD reactors and silane. It's not going to be easy to get away from explosive toxic gases and dirty processes.

kurthr 9y ago
Not to mention DiBorane and Arsene... 1000C anneal cycles, teratogenic photoresist ethers, UHV equipment, and muti-kW RF sources. It's possible to demonstrate um scale MOS with a small lab, but not safely in anything resembling a normal garage. Hoping to replicate a multi-$B process at home is a fool's errand.
- sevensor 9y ago
  While we're at it, a modern scanner costs many million dollars, weighs tons, and sits on its own separate foundation on pilings that go down to bedrock so that it doesn't pick up vibrations from the air handling systems. DIY it ain't :)
  - DoctorOetker 9y ago
    what is a scanner in this context? what is its role? or is this a stepper? or scanning electron microscope?
    - sevensor 9y ago
      A scanner is a photolithography tool like a stepper, except instead of exposing the entire shot at once, it sweeps (scans) a narrow rectangular exposure across the shot, and then steps to the next one. This allows for lighter, more precise optics. Confusingly, there's a piece of standard office equipment with the same name. Even more confusingly, Canon makes both kinds! Be sure you call the right technician for service.