There's been an interesting, if long, series of technically oriented comments on a recent post. A commenter named Mark Wendman has asked some technical questions about how molecular manufacturing (MM) might work, and made some rather skeptical assertions.
In one comment, he refers to "Positional drift, Tip durability / lifetime, positional verification, solid phase desorbtion source - stochastic emission and gas phase - beam induced polymerization scattering??" as possible problems. And in a later comment he expands on four of these points.
The discussion is interesting for several reasons. First, it shows how easily people can talk past each other, and even take offense. Second, it shows how a skeptic can criticize off-topic proposals without ever realizing it. This is an ongoing problem for MM, because there are a lot of things that won't work, and a skeptic who thinks the non-workable ideas are what's been proposed for MM will go away convinced that we're a bunch of shallow flakes.
Mark's objections deserve an answer, and rather than bury it at the end of a long comment thread, I (Chris Phoenix, CRN's Director of Research) will post it here.
First, some background, because Mark seems to be starting from some incorrect assumptions.
- Molecular manufacturing proposes to build a form of scanning probe microscope on a sub-micron scale. Early in the thread, there was some inconclusive discussion of scalability and exponential manufacturing that wandered into present-day diamond manufacture -- quite off-topic. I'm not sure Mark has realized that we are talking about such radically different and smaller SPM architectures. Since the time for an SPM to process its own mass in atoms decreases as roughly the fourth power of the size, a 100-nm SPM should in theory process its own mass in about 100 seconds.
- MM does not plan to build *every* substance or *every* product. However, it proposes to build a reasonable subset of highly crosslinked carbon-backbone covalent solids, in programmable shapes and with atomic precision. This should be enough flexibility to implement nanoscale machines (such as SPM's) and to implement a fairly wide range of micro- and macro-scale material properties.
- Probe-based MM (as opposed to polymer-based MM, a whole different topic that we're not even touching on there) probably depends on knowing where all the potentially reactive atoms in the system are. This is a stringent constraint, which may appear impossible in comparison to today's technologies -- but we can have a conversation about that. And if this condition can be achieved, it implies a lot less need for scanning/sensing and computation than exists in today's research systems.
Mark's first objection: Positional drift. In today's SPM systems, this is a major problem. There are several reasons for it, including properties of piezoelectric elements that will not be relevant, and thermal distortion of large systems that will be less relevant. (Thermal noise is not a source of drift, but instead a source of uncertainty. Any SPM system will have to deal with thermal noise, but so far it looks manageable, especially at moderate cryogenic temperatures.) Positional drift is a major reason why SPM systems need to re-register their position. A system with no drift should be able to skip that.
Mark's second objection: Tip lifetime. In today's systems that scan unknown and dirty surfaces with tips that are crudely structured at the atomic level, the tips gain and lose atoms pretty frequently. But would there be any wear if a buckytube tip were scanning a graphite sheet in a perfectly clean system? This is probably closer to the conditions that would exist in the hypothetical MM system. Mark, do you know if there is any experimental data on tip wear in clean graphene-on-graphite scanning? Perhaps the graphite-on-graphite superlubricity experiments, in which entire sheets pass each other with low friction (does this indicate absence of cross-linking?) might also indicate hope for low-wear contact.
Mark's third objection: Using the tip as the source of atoms will cause imprecision and tip wear. In today's systems, either the tip structure is the source of atoms, or the atoms (molecules) are adsorbed on the tip--not chemically bonded, but free to move around. In MM proposals, the tip would incorporate a precise structure containing one or a few atoms; would deposit those atoms, leaving a precise structure; and would then be recharged (have the atoms replaced) before the next deposition operation. No system today does this, and as far as I know only Freitas's proposal contemplates doing anything like this with today's technology. Deposition from a single precisely known molecular structure that's designed for that purpose can't be compared with today's deposition results. It should be criticized on its own merits.
Mark's fourth objection: Tip-mediated or beam-mediated deposition of atoms from gas is quite imprecise. I completely agree, but this has no relation to molecular manufacturing proposals. (It may relate to some bootstrapping proposals that are intended to make small, useful, imprecise structures. But it does not relate to post-bootstrapping exponential-manufacturing plans, which are the plans that deserve the most skepticism.)
Another objection that Mark made elsewhere in the thread was that surface-adsorbed atoms are mobile even at liquid helium temperatures. This is true, but MM does not propose to use adsorbed atoms. All the reactive atoms are intended to be covalently bonded to either the tip or the workpiece. Covalently bonded atoms are not mobile even at room temperature.
I hope that Mark will find these responses constructive. There are several responses he could make.
- "I refuse to believe that you can solve the practical problem of ______ so I will remain a critic." This answer ends the conversation. Possible ______'s include sub-micron SPM's (including actuation, but not including the control computer), contaminant-free workspaces, and doing carbon-lattice mechanosynthesis reactions reliably.
- "I don't see how you can do ______ (or be sure of ______) because ______ so you have to anwer that." If I get this response from Mark (or almost anyone else), I will answer it.
- "Oh, I guess I have to learn more about the actual proposals before I criticize them." Eventually, I hope to get Mark (and other skeptics/critics) to this position.
- "I can prove that ______, which is necessary to your plans, is impossible." I would be extremely surprised at such a response, because no one has found a showstopper in over two decades of trying, and some very smart people have tried.
- "You have to prove that _______ is possible before I will listen to you." Like the first response, this ends the conversation.
I hope that this conversation will be productive.