• Google
    This Blog Web

October 2011

Sun Mon Tue Wed Thu Fri Sat
            1
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

« Awakening Giants | Main | Say It Ain't So! »

March 31, 2005

Comments

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

cdnprodigy

Taking a few factors into consideration, considering the abilities of a few national entities to mount crash nano-programmes near the end of the next decades. This is my guess at the order in which they would succeed:
USA.
China. If 1st assembler is reached after 2020, good chance of 1st build here.
EU. Would be ahead of China and maybe USA, if it was certain national programmes would harmonize.
Japan.
India.
I think the top two have a distinctive edge over the others. I don't know if this tranlates into a lead of days or years. If Japan, USA, or Europe goes bankrupt, or if an avian flu pandemic hits China disproportionately hard, things will change. Whatever alliance forms, would likely have to include at least USA and China to be effective, if not the others. And this nano-Security Council would have to have an effective enforcement mechanism, unlike our current Security Council.

Philip Moriarty

cdnprodigy,

You may wish to visit Richard Jones' Soft Machines website for a rather more sober and, to my mind, realistic appraisal of the possibility of MM being developed in the near future.

The following links may be of interest:

Debating Nanotechnologies

Bits and Atoms

and

Is mechanosynthesis feasible? The debate continues.

Best wishes,

Philip

Philip Moriarty

I have made the following observations in recent posts:

1. It is unscientific to quantify arbitrarily the probability of the appearance of MM within 10 years when there is no basis for the quantification. (If this basis does indeed exist, I’d appreciate it if CRN could please explain the underlying reasoning because it eludes me entirely). The quantification is therefore entirely unwarranted and wholly misleading.

2. It is similarly extremely misleading for CRN to state that every issue ‘appears solvable’ and that “if you are able to look at the whole thing, you will see that the holes have already been thought of and filled” when a workable, experimentally viable scheme to carry out the most basic mechanosynthetic processes is still lacking. Moreover, many ‘holes’ in current MM strategies have been raised in previous debates and have either been entirely ignored by CRN or have not been addressed (i.e. solved) in any convincing fashion. (See also this post and this discussion .)

3. A solution is distinct from a set of untested ideas. [If CRN is confident that workable and robust solutions to the development of a MM technology have already been developed, the next logical step is to collaborate with experimental groups to realise (physically) these solutions. As an experimental physicist, I certainly have not seen a “solution” that might be implemented in the lab. Moreover, there remains no experimental programme of research devoted to the development of MM (i.e. molecule-by-molecule construction of covalently bound solids with computer-controlled atomic precision).] Simply stating "Designs can be found that will work" as Chris Phoenix has recently written in response to Richard Jones' comments, is not the same as providing a solution. If diamondoid is now to be neglected and sapphire to be used instead, where are the mechanosynthetic reactions and tools to synthesise sapphire? I reiterate: "speculation" and "solution" are not synonymous.

In putting forward these observations I have been accused of ‘snide rudeness’ on more than one occasion. Although irksome, I’ll ignore those accusations and leave it to the reader to decide re. the validity of the observations above. As noted here , there have been a number of interesting comments re. nanoscale thermodynamics posted on the CRN blog of late. I plan to comment on these later this evening.

Best wishes,

Philip

cdnprodigy

Yes, tools are generally cheap and widely dispersed. I just found out I can make an STM for a few hundred dollars and a few hundred hours labour; I'm gonna give it a shot. But I won't beat Intel to an MM. And any corporate or university effort will be dwarved by government efforts funded in the trillions. Nanhattan Projects would have access to more personnel in wider ranges of disciplines, would be better funded, and might even utilize classified technologies. "Garage MM" efforts would eventually work if an assembler is engineerable, but enjoy parity only in available computation resources. Cutting edge lithography tools are driving Moore's Law in the semiconductor arena, but tool prices are increasing exponentially. At current rates, only entities with tens or hundreds of billions of dollars to burn will be able to compete even without the entry of Nanhattans. I mention lithography because it seems the MM avenue easiest to quantify reliable progress. Also, an MM's computer components and heat radiation components should be commercially available thx to semiconductor industry.

Chris Phoenix, CRN

I'm not at all sure that lithography is the best approach to MM. It's inherently imprecise, and I'm not sure why a MEMS system would be better than larger scanning probes for building the first nanoscale mfg system.

About once a year I change my opinion on whether it's better to use scanning probes and go straight for diamondoid (or other covalent solid) construction, or go for a wet-chemical bootstrapping route. At the moment, wet bootstrapping is in the lead, because of recent development of better molecular building blocks.

Wet bootstrapping could probably be carried out in a smallish lab. One lab would still have trouble beating a Nanhattan project, because it would have to make its mistakes in series rather than trying several approaches at once. But a very smart researcher with a few million dollars of equipment and several hand-picked assistants might beat a politicized plodding billion-dollar effort that has no clear vision.

Chris

jim moore

"go for a wet-chemical bootstrapping route"

Wow that sounds like a really neat idea. One of the real interesting geometric consequences of becoming smaller is that you change the ratio between surface area and volume. At the nano-scale nearly all of an object is on the surface. So when we think about design issues maybe we can find a way to exploit the very strong surface forces. You could use hydrophobic- hydrophilic interactions to provide structure and orientation for an object. If you can change how much a section of an object "likes" being in water, you can get that section of the object to move in relation to water. With nano-scale objects constantly being bombarded by molecules of water, maybe we can use a design that takes advantage of the vibrations and rotations that are induced. That might mean we look at introducing unusual levels of flexibility into nano-objects. (the objects/machines might even be "soft") We can "steal" a lot of design ideas from biology, we might even take some molecular components directly from biology.

But if start down this road we will not be able to use a macro scale mechanical engineering paradigm to design wet nano-scale machine systems. The physics of these types of nano-scale objects/ systems can get pretty complicated. For example getting a real quantitative understanding of the effects of entropy is something that eludes me. (My understanding is more qualitative, highly ordered states are unusual, randomly ordered systems are more common)

Hrmmm. You know, it would be pretty cool if CRN had contributor who knew a whole lot about this wet chemistry approach to nanotechnology. We might be able to chart out a fairly detailed road map for "nano building nano" using this wet chemistry approach.

Chris Phoenix, CRN

Jim, I think you've rediscovered the mainstream approach to building nanoscale systems. Yes, biology has done some pretty amazing things, given the materials it has to work with. And yes, it uses applications of physics that we are not very familiar with.

But biology has to evolve, and metabolize itself, and live in water. Human-built nanomachines don't have to do any of that. There's a huge design space that biology couldn't access.

A system with too few degrees of freedom in its operation can't evolve; change a component randomly, and the whole thing breaks. I suspect that most of the complexity of biology comes from that.

You can't use thermal vibrations in the sense of extracting energy from them. You can sometimes use them to lubricate motions; this is equally true for stiff and soft machines.

You can use surface forces to make things stick together. Using this for Brownian assembly ("self-assembly") has two problems. First, the forces have to be weak enough to back out of bad configurations. Second, all the information has to be embodied in the molecules. If the assembly patterns/motions are externally supplied, then surface forces can be a useful glue.

So... when I said wet-chemistry bootstrapping, I was not thinking of trying to emulate biology's floppiness or complexity. I have a design in mind that uses self-assembly to attach ~5-nanometer stiff molecular building blocks to manipulators, then uses the manipulators to attach the blocks strongly to the product. (This can be done with positioning, with optional addition of pressure, light, chemicals, or electricity.) It should be possible to build a simple manipulator this way. And the first manipulator can be a scanning probe microscope with a functionalized tip.

Chris

Chris Phoenix, CRN

Jim, on your last comment, "You know, it would be pretty cool if CRN had contributor who knew a whole lot about this wet chemistry approach to nanotechnology. We might be able to chart out a fairly detailed road map for "nano building nano" using this wet chemistry approach."

I think the wet-chemistry people will be the first to tell you that no one knows a whole lot about how to do a wet chemistry approach to nanotech. They've only recently figured out how ATP synthase works. They could not design a ribosome. (The functionally closest thing we have is Seeman's programmable DNA-building machine, and it uses stiff structure and simple motion.)

I have talked some with Robert Bradbury, who likes the idea of engineering bacteria. This is more biotech than nanotech. It seems a good way to make lots of chemicals from a limited set. So far, I haven't seen any proposals for how to make large structures, or fast digital logic, etc.

Basically, the "hard machine" nanotech people can design things they can't build, and the "soft machine" nanotech people can build things they can't design. But building stuff is an engineering problem and can be bootstrapped. Learning to design stuff with physics we don't yet understand seems likely to take longer.

And in the end, a dry system will have much higher performance than a wet system, simply due to drag. Shallow design methodologies will waste some of this performance, but we have literally orders of magnitude to burn. And dry systems can use many of the physics tricks that we attribute to wet systems, so as the physics and design skill develops, both camps will be able to take advantage of it.

Yes, there is probably a roadmap waiting to be identified that uses protein and DNA engineering to build better ribosomes and augmented proteins and bootstraps from there. In fact, Drexler has always argued that this is the way to go. One of the Feynman prizes last year was for protein structure design; they invented a new fold and it worked. That's a ways from engineering a ribosome, though.

I expect that it will turn out that, in non-evolving systems, the best machines for a job are ones where the degrees of freedom in the machine's operation are minimal. Certainly that will be easiest to design. And in stiff machines, there's no need to use lots of DOF to do efficient state transitions.

Hm--an insight. I think. I said that the complexity of biology was necessary for evolution. That's true. But I suspect it has another purpose as well. As a biomachine moves, in order to be efficient, it has to balance forces and avoid any steep force gradients. By putting lots of dangly squishy bits all over the molecule, the entropy of each part of the trajectory can be fine-tuned. And this "variable entropic spring" can balance the forces along the trajectory.

For the same purpose, a hard machine could simply use a mechanical spring and a cam.

Chris

Chris Phoenix, CRN


OK, I've done some more thinking on the entropic spring, and realized a few more things. The writeup below explains why tortuous motion in biomachines actually makes design easier. It also points out that entropic-spring designs have to move slower than mechanical-spring designs for the same efficiency.

Bio-motors can be extremely efficient. This implies that they have good control of energy over a very complex trajectory. This seems near-magical.

Start with a machine that has an inefficient and tortuous trajectory, with varying unbalanced forces at various points. Now add dangly floppy bits to the machine. As the machine moves and changes shape, the entropy of the floppy bits will change. This acts as a variable entropic spring. And the floppy bits can be modified evolutionarily to fine-tune the entropy to balance the energy (path-integrated force).

The point of the tortuous trajectory is to provide enough physical configuration change in the machine that the entropic spring is easy to tune. The more complex the trajectory, the more freedom evolution has to optimize the spring.

If you don't have to use an entropic spring, then you can use a much simpler trajectory. A stiff machine could do force-balancing by linking the motion to a mechanical spring following a cam. For this approach, simpler trajectories will be easier to engineer.

The entropic spring requires fairly slow motion, so that each new configuration has time to access its whole space. Otherwise, springiness will be lost. A mechanical spring-and-cam should have much lower drag at any given speed because it has a much smaller volume (both physical and configuration space) to equalize.

Chris

jim moore

Chris,
A couple of points;
1.) I think that the "entropic spring" vs Cam and mechanical spring is very interesting idea and I would love to see a computer simulation (done at room temperature and in water) of the two designs. Your cam and spring design sounds a lot easier to design but I would really like to see a "working" model.

2.) I think that you may have missed the real message I was trying to convey in my first post. In a gentle and non-confrontational way, I was trying to get you to realize that one of the most likely pathways to diamondiod nanotech goes through an area of expertise of a certain polymer physicist that you told to go away. And that maybe, you would come to the conclusion that you made a mistake by doing so. I was hoping that if you realized that you made a mistake that you would try and repair a relationship that has the potential to help you reach your goals.

Chris Phoenix, CRN

Jim, cam and spring designs should not care whether there is water present or not; at most, they'd have to be tweaked a bit. (Assuming that there were no chemical instability problems from the water.) The entropic spring approach will work with or without water, but I'd expect it to need quite different parameters on the entropic parts. And of course if the structure required water (e.g. hydrophobic/hydrophilic) then it wouldn't work dry.

As far as I can see, the only thing you need for a cam and spring design is a low-friction interface between the cam and the spring, and a linkage that is stiffer than the force it's trying to balance. It's a very general technique, and the only reason to simulate it would be to test a particular atomic-detail implementation.

I will use anyone's expertise who doesn't try to confuse me. There's even some value in being confused; I've gained several insights in the process of figuring out why certain objections don't apply to the domain of stiff room-temperature covalent-solid machines. But I've also annoyed some of my experts when I went to them for help in answering bogus objections. And I can't simultaneously do science and debating. If you know of any suitable experts, please put me in touch with them.

Chris

jim moore

"It's a very general technique, and the only reason to simulate it would be to test a particular atomic-detail implementation."

If the cam and spring mechanism really can work at the nano-scale, a good computer simulation of a particular design working would be very educational for nano-scale designers and engineers.

Would it be possible to make a working nano-scale cam and spring mechanism out of DNA or RNA? Or is it only doable with diamond and graphite?

Chris Phoenix, CRN


Was the simulation of Drexler's planetary gear educational for nano-scale designers and engineers? Or was it just a target for skeptics to say "Look at them making pretty pictures of stuff they'll never be able to build"?

I don't know if it could be done with DNA. I've never seen a sliding interface done with DNA. Hm... you could pull something based on a zinc finger or a restriction enzyme along a DNA strand, and use the sequence of base pairs to change the energy. Or you might build a variable-force spring based on unzipping complementary strands that have a few base pair mismatches. (Will DNA zip around a single mismatch? I think so but I'm not sure.)

But there's a whole range of materials between DNA and diamond, starting with proteins with varying degrees of crosslinking.

If there's a a nano-scale designer out there who wants an educational experience, doing the simulation would be a much better experience than seeing the simulation. Compute the required force for the application you'll be using the force compensator in. That will tell you what tradeoff needs to be made between material stiffness and scale. Pick a material that fits your scale requirements and that you can implement a sliding interface with. Then do the mechanical design and make a movie.

Chris

jim moore

Was the simulation of Drexler's planetary gear educational for nano-scale designers and engineers?
Yes

Or was it just a target for skeptics to say "Look at them making pretty pictures of stuff they'll never be able to build"?
Yes, it was that as well.


I think that whenever anyone attempts to do something in an unorthodox manor they open themselves up to ridicule. If the unorthodox way turns out to be better, victory is so much sweeter.

Philip Moriarty

Mike, Chris,

Your recent changes to the manner in which hyperlinks are treated in comments on the CRN blog has removed all the links in the posts above (and elsewhere on the CRN blog).

Philip

The comments to this entry are closed.