• Google
    This Blog Web

October 2011

Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          

RSS Feed

Bookmark and Share

Email Feed

  • Powered by FeedBlitz

« CRN at Five Years Old | Main | Debating CRN's Scope »

February 05, 2008


Feed You can follow this conversation by subscribing to the comment feed for this post.

Tom Craver

Vince -

First, my "make it bigger" approach is a bit tongue in cheek - my main point is that all of these are *engineering* challenges, while Richard presents them as if they are reasons not to get started.

#1 is about reconstructing in products - larger features makes that less of an issue. There are dozens of possible ways around the reconstruction issue during building - just change your assumptions.

For #2 - brownian motion energy and thermal energy clearly DO get easier to handle with larger sizes - else factory floors would be full of machines that bounce and flop all over the place. Inertia and rigidity grow as cube and square of size, while thermal effects start to cancel out. Yes, "how big is big enough" is still an open question.

#3 - Again, yes, "how big" is still open. I'm not saying these aren't "challenges" - but it looks to me like Richard presents these as "don't bother trying to develop hard nanotech yet - not enough is known, and it probably can't work".

But that's exactly backwards - focusing on engineering 'hard' nanotech is the most efficient route to finding the solutions - solutions are almost certainly possible to all of these, given the huge range of approaches. I'm sure there are more subtle approaches than "make it bigger" - but that's really not a bad place to start.

#5 - Making devices bigger reduces the need for a perfectly eutactic environment. E.g., by operating at a slightly larger scale, you might be able to use a very low pressure flow of a cool, relatively inert gas, to sweep out less desirable molecules. You could maintain an outer shell maintained at greater than ambient pressure, so that few pollutants ever get past that shell.

So many potential solutions to such a small number of challenges! :-)


Brian, your arguments about the Wright brothers and railway engineering are inspirational and show a positive attitude on your part. You also describe useful general strategies for solving engineering problems such as breaking them down into simple steps. But sometimes technical problems require technical solutions. The strategies you present are not solutions to the challenges. They are ways of working towards answers.

Here's another analogy to think about. Imagine you are chief engineer of a semiconductor processing plant and one of your engineers approaches you and tells you that the 32 nm node process is suffering from electromigration faults, excessive leakage currents and low yields. Would you say to her "These are merely engineering problems. Work around the electromigration problem, reduce the leakage currents and as for the yields, make it as good as you need to and improve later if you have to"? If you did she would not ask you for help again.

The 6 challenges do not have simple answers right now. These are some of the big questions that Freitas and Zyvex will almost certainly be thinking about. More calculations and experiments may eventually find solutions to all of them or prove them to be irrelevant. This could take months or it might never happen. Noone currently knows. This is why the CRN and Zyvex timelines failed.

Don't get me wrong. Its a good idea for you to try and think of solutions in blog posts and it is good that this forum exists to allow people to exchange ideas. But CRN have to be prepared to defend their arguments and accept criticism when others point out flaws in their reasoning.


Tom, sorry, I missed your tongue in cheek meaning. But I am sure you have misunderstood Richard's post and I think you should take another look at it. My reading was that his intention was exactly the opposite of what you said. The challenges are presented as research questions to instigate more focussed activity and discussion.

Both yourself and Brian talked about these being engineering challenges. It doesn't matter how you classify them, they still need to be solved.

On #1, #2 and #5 I would agree with you that these issues may be less important if the components are an order of magnitude larger but only if you allow design tolerances to scale up as well. The position of CRN and Drexler has always been to insist on atomically precise structures. These challenges remain as stated if this atomic precision tolerance is required as a design rule.

For #1 what are the dozens of way around the reconstructions? Given that Freitas has only published a few papers on a very simple mechanosynthesis step how can you know this?

For #2 the problem is achieving atomic resolution when each linkage in the mechanism of your robot arms has a finite positioning error associated with it. These will compound. Thermal effects do not cancel out at the nanoscale. That would defeat the laws of thermodynamics.

For #5 an inert gas will slow down the moving parts and generate a lot of heat. That is why it has been ruled out by mechanosynthesis proponents. Impurities will enter the system with the feedstock, that is where the filtering is needed. A pressurised outer shell will not help.

Tom Craver

Vince -

You need to understand that Richard is promoting an alternative "soft nano" approach - for which I say more power to him, except that he seems to feel that he has to "destroy the competition" in order to promote his ideas.

#1 - One assumption is that one atom at a time will be added to an incomplete crystal face. Another assumption is that atoms must be put in place correctly once and must stay there - that subsequent adjustments can't be made. Again, I'm not saying "I've got the solution" - I'm saying "broaden your perspective on possible approaches". If atoms are going to move after being put down, you might think in terms of a "jig" - a temporary molecular layer that fills the area around the assembly site and counters the forces that cause reconstruction. Or lay down an entire surface of atoms, then make additional passes to push atoms into the correct bonding structure. Get creative and stop thinking about fat fingered robot arms pressing atoms into place one at a time.

#2 - And yet, AFMs work, despite having tips on the atom scale. At very worst, a 1st generation nanofactory unit could be attached to the tip of an AFM to achieve sub-atom positioning accuracy. A second generation might shrink the positioning equipment down as far as makes sense, and no further. Subsequent generations might find clever engineering solutions, based on experience and study of real nanomachines, to shrink it even further.

#5 - You seem to be working from the idea that there is only one, ideal nanofactory, and that we must build that ideal right from the start. The first "real" nanofactory unit is likely to be slow, over-sized, crude, run hot, quickly break down, etc. It's primary virtues will lie in providing experience, and easing the process of building a next generation, which in turn will enable another and so on until we get something that works well enough to go commercial - and even that won't be the end of it. It does not *matter* if version #1 runs slow and hot, as long as it runs. This is why I keep harping on taking an engineering perspective, rather than a theoretical science perspective. Pointing to problems is useful - but at some point you have to start searching the solution space if you wish to achieve anything.

And btw - the pressurized outer-shell approach was in response to the issue of needing to mechanically affect things outside a nano machine, which is part of what Richard describes in #5.



If you read Freitas, Merkle et al papers then you would know one of the answers to avoiding reconstruction. Note: I also mentioned this work in the comments to the original 6 challenge thread to Richard

The Germanium version of the DBC6 dimer placement tool had detailed computer simulations performed showing that it should work reliably at room temperature. There is no problem placing dimers with a one dimer gap. Placing them directly adjacent can cause some defect formation.

The paper also describes using a first pass of placing every other dimer. Then adding the skipped dimers. This process is a defect-reduced procedure for fully populated dimer rows.

The simulations also showed the maximum placement tolerances. 0.5 angstroms in X (or across dimension) and 1.0 angstroms in the Y (or along trough) for an isolated dimer.

1.0 angstrom in the X and 0.3 angstroms in the Y for a correct gapped dimer placement.

Richard's responses were:
it should be combined with a programme of experimental validation to make progress on my challenge 6. [Which Freitas and Merkle are doing to achieve their goal not necessarily for his challenge]

In specific response to Brian, the point in challenge 1 is much more general than whether this or that mechanosynthesis step will work. It is that, currently, we do not know whether the structures being drawn as targets for the mechanosynthesis - the rotors and cogs and gears - are chemically stable under any given set of conditions, let alone whether the necessary intermediates are. Chemical stability is about much more than just satisfying the valency, and none of us are able to do density functional theory in our heads.

Which is why top researchers and engineers have to put in the work and solve all the problems. Tossing 6 challenges over a wall and waiting will not result in solutions.

So more resources (money and people) should get steered to working on this because the prize is clearly worth the effort to try to succeed.

As many mothers say, the room will not get cleaned and will not happen today if no one works on the chore. If no one works on MNT between now and 1 million AD then it will not get done by that time. But if people do work on it then more problems will be solved and it will happen sooner.

Part of the optimism in CRN timeline assumptions was that at some point the tide would turn and a lot of effort ala the Manhattan project would be thrown at the problem. There was some thinking that what is now the NNI would have been that effort but instead those billions are directed at relabeled chemistry projects and other existing work that is now packaged at a broad definition of nanotechnology.

Mike Treder, CRN

You're right, Brian, about CRN's expectation that the "tide will turn" at some point and much more funding and effort will be put into developing MM. It hasn't happened yet, and maybe it never will, but if it does, our worry is that the science could proceed quite rapidly (given sufficient support and good management of resources), and may in fact achieve a breakthrough well before adequate steps have been taken to prepare for the technology's disruptive potential.


Brian, the paper you mention is about putting carbon dimers on C110. If it worked in practice then you could add carbon to a flat 110 surface in layers of decreasing size. To make anything useful you need to be able to make edges and surfaces that go at other angles. Noone has claimed they can do this yet. The paper doesn't address the problem of how to make any of the parts in the nanoengineer gallery either. It marks the start of a long process of research. Lets be clear on this. We are still at a very early stage.

"Which is why top researchers and engineers have to put in the work and solve all the problems. Tossing 6 challenges over a wall and waiting will not result in solutions."

I agree with you that the problems are unsolved and will remain unsloved until people solve them. We don't know if or when this will happen. That is why CRN is wrong to say that the goal will almost certainly be achieved by 2025.

It is interesting that you compare Richard posing his challenges with someone tossing a missile over a wall. Do you think that the effect of his simple questioning was to cause destruction? Is mechanosynthesis founded on such fragile arguments? This was your analogy, not mine. Would you prefer we lived in a world where ideas were not criticised objectively?

"So more resources (money and people) should get steered to working on this because the prize is clearly worth the effort to try to succeed."

Your argument is fallacious. As I have explained, the effort in time and money needed to succeed is unknown. Take a look at Zyvex for example. Some of the keenest mechanosynthesis experts in the world completely underestimated how much time and money was needed to make progess. They still don't know how long it will take. If the effort needed is unknown then you cannot say it is worth making the expenditure. Funds are limited and many other worthwhile research areas are competing for them.

Anyone who wants to propose research in mechanosynthesis can compete for funding in the west. They always have been able to. But we don't give away money. Proposals are examined critically. CRN's wild and unsubstantiated claims have more potential to hinder responsible publicly funded mechanosynthesis research than to help it.


I did not say "toss a missile over a wall". I said toss it over a wall referring to the common business phrase of tossing a business idea over a cubicle wall.

Here is a link with an explanation of this common phrase "throw it over the wall"

To pass a project or problem to another person or department without consulting with them or coordinating the transfer in any way.

Someone in say marketing writes a note on a piece of paper, crumples it up and tosses it over the wall [over to another department like research]. The phrase indicates that it would be ridiculous for someone in marketing to then expect the other department of the company R&D to do something useful with the suggestion on the note.

It does not indicate that the note is destructive but that it is relatively useless.

I am saying that this behavior shows a lack of interest in solving problems and achieving project success.

There is plenty of substance behind the claims that CRN has made. You are choosing to belittle it or ignore it.

You are hopping around from technical feasibility to project cost estimates.

>Anyone who wants to propose research in mechanosynthesis can compete for funding in the west.

You are wrong the competition for funding for mechanosynthesis is not based on unbiased rational competition.

Before certain interests got it removed the NNI billion dollars per year in funding had 5-10% of the money assigned to projects to rigorously prove or disprove the feasibility of molecular nanotechnology. We would not have to still be guessing a decade later if 50-100 million per year was put into trying to make this work and see what was really possible. International nuclear fusion has gotten more money for a longer term project. There was also talk about putting similar money aside for exploring safety of nanoparticles or on nanoethic issues. Why were the Richard Smalley's and Mark Modzelewski's of the world so scared of that they did not want to check to see if this might work ? Despite the revolutionary promise of molecular nanotechnology (MNT), the U.S. National Nanotechnology Initiative (NNI) had explicitly excluded funding for MNT.

This article by Ed Regis explains the politics of how MNT did not get funded. It was not because of the basis of science

After years of lobbying by the Foresight Institute, in May 2003 the House passed the Nanotechnology Research and Development Act by a lopsided vote of 405 to 19. The bill contained a provision - written by California representative Brad Sherman, a Drexler supporter who had spoken at Foresight's annual conference the previous year - calling for a study to "develop, insofar as possible, a consensus on whether molecular manufacturing is technically feasible." If the technology was deemed feasible, the study would find "the estimated time frame in which molecular manufacturing may be possible on a commercial scale; and recommendations for a research agenda necessary to achieve this result."

With this language, Congress was on the verge of making Drexler's dream a reality. But by November - five months later - the provision had vanished from the legislation.

Molecular manufacturing is a "loaded term," a Senate staffer says. "It upsets a lot of people."

The sponsors of the House bill were more interested in making sure it got through the Senate than they were in preserving funding for Drexler's ideas. Thus, when House and Senate staff members met to discuss their respective bills, they scuttled the molecular manufacturing study. In the Senate version, Arizona's John McCain introduced an "amendment in the nature of a substitute" in which the provision no longer appeared.


Brian, I apologise for misunderstanding your analogy. I wasn't aware of the phrase as you described it. But did Richard really pass his challenges on to anyone to deal with without consultation? He posted them on his own blog for the world to see. He did not ask anyone in particular to solve them. Your analogy is unfair.

I think it is entirely correct that someone is entitled to post constructive criticism of someone else's ideas. And as it says in his post, the purpose of the challenges is to focus attention on the problems which need more research. This shows genuine interest in solving problems and achieving project success. I completely disagree with you.

Hopping around?

I am not hopping around. I am criticising the unsubstantiated timeline of 2025. I am supporting my argument for this by pointing to a list of 6 genuine unresolved challenges which have received scant attention in 2 years. This shows how slow progress is, how much remains to be done, and therefore how unrealistic the timeline is.

Lets be clear, I am only addressing technical feasibility points raised by you and Tom because you are claiming that the challenges are solved or easily dealt with. They are not.

"There is plenty of substance behind the claims that CRN has made. You are choosing to belittle it or ignore it."

There is no substance behind the 2025 claim. Where is there a breakdown showing how this will be achieved? Where is the gantt chart?

On funding

I stand by my claim that there is money available for research. Zyvex's APM funding is proof. Do not get stuck on one failed proposal from years ago. Make your proposal better and try again.


Tom, I was surprised to see you quote the words "destroy the competition". Where did you get that from? I refer you to my response to Brian above. If a fair analysis of CRN's position results in the destruction of their arguments then does that not suggest to you that their arguments should be treated with suspicion?

#1 Where did you get these assumptions from (one atom at a time, subsequent adjustments can't be made)? 1. Freitas proposed desposition of two atoms at a time. This is well known and is not an assumption made in the post. 2. Richard talked about the possibility of transforming metastable structures into useful configurations. He explicitly said that subsequent adjustments could be made. As far as I can see the assumption is one you have made, not him. Again, did you read the post?

Your suggestions about making a jig out of other molecules are helpful. Maybe research in this direction will work, and maybe it won't. The same applies to your idea about pushing atoms around.

#2 AFMs do not work in the same way as the nanofactory robotic mechanisms. They use large piezoelectric crystals. This is a completely different technology and it doesn't scale down to the nano-scale. Lets be clear here. The nanofactory and mechanosynthesis proposals talk about robotic mechanisms. Challenge #2 is potentially a big problem for these.

#5 Again you are inventing assumptions. I do not assume that a nanofactory has to be built ideal from the start. I am only providing criticism of the proposed final design. You suggested using a low pressure gas as a way of flushing out the impurities that would cause the problem of contamination raised by Richard. I explained why a gas cannot be used. The challenge remains.

The challenges are a step forward for mechanosynthesis because they set out research which can be carried out now. You can suggest ideas to work around them in a blog, but until the ideas have been demonstrated in technical detail in theory or experiment then the challenges remain unsolved. The reason I mentioned the 6 challenges here was to show how in 2 years CRN have refused to acknowledge these difficulties and reflect them in their writing.

As I wrote earlier, the claim that "On the point of whether or not molecular manufacturing is feasible, CRN and our allies apparently have won the argument." defies explanation. The 2025 timeline claim is also completely substantiated. The more time that passes without CRN addressing these problems the more ludicrous their position appears.


1 billion per year in research funding that is labelled Nanotechnology is specifically not available for open competition for molecular nanotechnology research projects.

The hijacked NNI funds come up year after year. It is not one proposal but the entire segment of funding which is cutoff to all proposals.

>posted it on his blog for the world to see

This is akin to putting a project critique on the company lunchroom bulletin board.

Actually trying to be part of solving the problems would be follow up postings and work devoted to determining how to avoid problems or how to make progress on the issues. Two years after the posting where has Richard actually tried to come up with any solutions ? He has influence over science and research budgets in the UK, has he come up with any projects to characterize issues more exactly or to make any progress ? No. So I stand behind my statement that these were scribblings that were displayed with no effort made on the part of the author to try to work at resolving them. Not only was no effort made to resolve them, when it is pointed out how some of them do not need to be resolved or where significant progress has been made there has been no update or clarification to his list. Before this list, Richard Jones made comments on this site that a particular molecular engine would not function because it would overheat. His same argument could be made to combustion engines that were solid. He ignored the basic concept of spacing and air gaps. I believe this revealed his bias towards only finding ways where diamondoid MNT will not work. As his book states he only believes in Soft machines.

So who is Richard working with to advance the area around the 6 challenges ? If he could definitely prove that MNT is not possible then why has he not gotten a peer reviewed research paper out of this area ?

So if someone was working in this area, why would they be doing better consulting Richard Jones on it ? When I drive someplace should I always pick up a passenger who says here are 6 reasons why we will not get where we are going. Are we there yet ? You suck ? I might ask "Seriously dude- are you going to spring for a tank of gas or help with the navigation map ? You know actually help get us there or find a good alternative destination that is close ?"

Plus I have already pointed you to the more comprenhensive list of issues that still need to be completed for a nanofactory compiled by Robert Freitas which is a superset of what Richard talks about. Plus Robert and Ralph are actively working to overcome all of them. You should look at the follow on work since the dimer placement work with the complete basic toolset that they characterized.

There is as much substance behind the 2025 estimate as any other projection of the arrival of significant future technology or future event. Like say Richard Smalley's
vision for the Armchair quantum wire was the eventual development of a distributed store-gen grid for 2050.
Would a gantt chart make the 2050 date more credible ?

>Lets be clear, I am only addressing technical feasibility points raised by you and Tom ..
Yes, I know you are trying to prove via argument that this is all wrong.

The comments to this entry are closed.