Re: Tip reactions... does Smalley have a point?

From: Chris Phoenix (cphoenixNOSPAM_at_crnano.org)
Date: 10/30/04 

Date: 30 Oct 2004 17:50:51 GMT

This is a very thoughtful discussion. It's refreshing to see thoughtful
inquiry after years of careless skepticism. Thanks!

I won't pretend that I have a satisfying answer. I have several
hand-waves that may improve the picture. I'm not a chemist, and I
haven't hashed this out with Drexler, so there are probably additional
arguments that make the project look more doable. But note that these
arguments don't all have to work--they're parallel options--and the best
answers are at the end.

(Advice for finding workarounds yourself: Don't start by looking for
reasons why it won't work. Look for ways to make it work. Once you've
found them, then criticize them. Similarly, the advice from Houdini:
When trying to figure out a magic trick, don't try to figure out how the
other guy did it; try to figure out how _you_ would do it.)

Handwaves below...

Jeremy wrote:

> John Larkin <jjlarkin@highlandSNIPtechTHISnologyPLEASE.com> wrote in message
> news:<cls8um02avd@enews3.newsguy.com>...
>
>>On 28 Oct 2004 18:50:25 GMT, lord_psi@hotmail.com (Jeremy) wrote:
>>>In the process of constructing the fine motion controller, there are
>>>going to be times when a site is dehydrogenated immediately adjacent
>>>to a surface bound oxygen atom, or some other reactive moiety.

Hm. Even if you're restricted to adding atoms one at a time to a
mostly-flat layer, then when trying to build a 1-nm cube (~100 atoms)
you have still have about 10^42 choices as to the sequence of atom
deposition. Of course most of these choices won't work. But if one
does work, and if you can find it, then you're fine. Is it possible
that the oxygens could generally be deposited last in their layer?

Other options: maybe a small set of tooltips can build a wide variety of
oxygen-containing molecules that can be deposited in one piece, so you
never have to have a reactive oxygen floating around.

>>>With sufficient mechanical energy, we
>>>shouldn't necessarily have to mimic enzymes ability to catalyze the
>>>reaction by binding to the transition state, but we still need
>>>stabilizing moieties surround the dedicated reaction tip without
>>>affecting the reaction. This is not a simple tip.

I agree, stabilizing lots of nearby reactive atoms using a non-bonded
tip seems quite difficult. Here are a couple of tongue-in-cheek ideas
from a non-chemist.

1) Before you insert the reactive molecule for deposition, plunk down a
ring of some weakly-binding atom like lead. Anything loose and reactive
on the surface will form weak bonds to the passivation ring. Now inject
the moiety to be deposited through the center of the ring. Now pull
away the ring, and if you haven't stirred things up too much, the
surface will be unchanged.

2) Pin down reactive floppy atoms temporarily, by oxygen-bonding them to
something out of the way. Peroxide bonds appear strong enough to
survive at room temperature, but weak enough that it shouldn't be too
much trouble to remove the pinning atom when you want to bond something
to the pinned atom.

>>>This may already have been addressed, and if so, can I get some
>>>references? If not... are we arguing past Smalley and his ilk? Yes,
>>>diamond should work. But can we make an assembler out of only
>>>diamond?

Actually, it's expected that the answer is yes. At least according to
Freitas and Merkle. Freitas even put up a chemical formula for his
assembler architecture: something like C400,000 H200,000 Si30 Sn30 (I'm
making up the numbers).

>>>I would encourage the modelers out there to construct the
>>>fine motion controller step-by-step, or at least model the critical
>>>geometries, including flat surfaces, corners, and edges.

I strongly endorse this suggestion!

>>I work with tomographic 3D atom probes, which *disassemble* things one
>>atom at a time. The samples have to be cooled to numbers like 20-40K,
>>because every time you remove one atom, the neighbors all want to move
>>around, and that messes up the structure you're trying to analyze.

But this is great news! If we can prevent rearrangement by cooling to
20-40K (which isn't all *that* cold) then it looks like we're home free.

> I've seen lovely
> models of carbon mechanosynthesis, apparently indicating that at room
> temperature such shifts aren't a problem when a carbon dimer is placed
> on a reconstructed surface,

Um, do you mean a depassivated surface? The Freitas models I've seen
use the 110 surface, unreconstructed.

Chris 

-- 
Chris Phoenix                                  cphoenix@CRNano.org
Director of Research
Center for Responsible Nanotechnology          http://CRNano.org