(University of Texas at Dallas) These unexpected but highly useful properties could have applications such as making composites, artificial muscles, gaskets or sensors. The team's findings are reported in the April 25 issue of the journal Science. (Source: 
(Manish Chhowalla, Rutgers University) Thin films of graphene could provide a cheap replacement for the transparent, conductive indium tin oxide electrodes used in organic solar cells. They could also replace the silicon thin-film transistors common in display screens. Graphene can transport electrons tens of times faster than silicon, so graphene-based transistors could work faster and consume less power. (Source:
At the 2007 Foresight Vision Weekend, Ralph Merkle and Robert Freitas in a joint session discussed research toward diamond mechanosynthesis. The two co-founders of the Nanofactory Collaboration offered a preview their recently accepted paper “A minimal toolset for diamond mechanosynthesis” to be published in JCTN (Journal of Computational and Theoretical Nanoscience). The talk included a description of nine molecular tools and all reaction pathways involved in their synthesis from raw materials. This took the co-authors three years to finish, and they view it as the critical next step in developing molecular nanotechnology for atomically-precise manufacturing.
The following transcription of the Foresight Vision Weekend presentation by Ralph Merkle and Robert Freitas entitled “Diamond Mechanosynthesis: A Direct Pathway to Diamondoid Nanofactories” has not been approved by the speakers. Video and audio are also available.
Diamond Mechanosynthesis
Ralph Merkle: The first thing I would like to do is simply say there are some web pages you could look at: molecularassembler.com. It’s got a bunch of information on it describing what is involved and how it is we move forward to make nanofactories and so forth and so on. It discusses this long-term goal of designing and ultimately building a diamondoid factory. Once you go to that domain name, then you can follow links around. One of them is on the nanofactory, which discusses various aspects of that, and links to other things that are of interest.
How many people have seen this slide? You’re joking. I’ve been using this slide for a decade now. All right, we have a new, fresh slide, which has never before been seen.
The basic idea is we’re talking about health, wealth and atoms. If you rearrange the atoms in coal, you get diamond. Those are the upper two pictures on the left. If you rearrange the atoms in sand and add a pinch of impuritie, then you get computer chips. And finally… I think I softened that picture, actually, to show someone in a hospital as opposed to someone dead. But as you can see the atoms involved in someone who is dead and someone who is alive and healthy are in fact the same atoms, and it’s how they are arranged that makes a big difference.
It’s very important how atoms are arranged. That is sort of the high-level discussion.
We’ve been seeing three major trends in manufacturing which are towards greater flexibility in the manufacturing process, greater precision in the manufacturing process, and lower cost. We’ve been seeing these trends over time, and as we look into the future, we’re going to reach a point where the limits of those trends are limits that we are actually approaching very closely, where we are actually able to arrange atoms in most of the ways permitted by physical law.
This is what Feynman was talking about in 1959. How many of you have read There’s Plenty of Room at the Bottom? Good, goo, good. Anyone who hasn’t, it’s up on the web. It’s an excellent read. It’s vintage Feynman; it’s great fun.
Now, when you are talking about arranging atoms the way you want, the first thing you say is, there are a hundred elements in the periodic table. Let’s throw out most of those and focus in on diamond. Diamond is really good stuff. Diamond has better materials properties than almost anything else. Diamond is hydrogen and carbon, and the carbon is mostly used to build the body and the hydrogen is used to terminate the surface.
If you look at diamond, it’s wonderful. It’s strong, it’s lighter, it has a bigger band gap, it’s more transparent, it has higher electron mobility. You name it, it’s got it. It’s usually either the best or close to the best in all the materials categories with maybe one or two exceptions.
Audience: Why is bigger banding gap good?
Bigger banding gap is good because it lets you have a higher operating temperature, so you can run this thing scalding hot. That’s a wonderful property. Scalding hot means it will tolerate a hotter environment, and you can run more power through it and it still works. The other reason it’s a useful property is that the smaller the banding gap, the more you have to worry about thermal noise. In other words, things bounce back and forth from one to the other so you have risk. If you’re talking about future electronics, a one-electron system, then it’s very nice if the one electron moves around and you don’t have thermal noise creating another electron or making an electron go away. It becomes an issue there. Wide band gaps are nice things to have. You also get a lot of other properties as well, in terms of diamond.
Audience: What about the fire hazard?
The fire hazard? Oh yeah, it burns. All right, diamond is not perfect. Various things are more fire resistant. You don’t want to use diamond for everything. For example, for the lining of the throat of a rocket.
Robert Freitas: That’s why we say “diamondoid” instead of “diamond.” “Diamondoid” includes sapphire.
Ralph Merkle: Of course, we’ll be mostly focusing mostly on the hydrocarbons at first, but the broader picture is you can include other stuff. Mostly I think of operating at low temperatures. Boiling water is fine.
What kind of things might you build if you had hydrogen and carbon? Here is a hydrocarbon bearing. It is basically two pieces of diamond that are bent. You take a piece of diamond, bend it into a collar, and take another piece of diamond, bend it into a bigger collar, stick the smaller one into the bigger one, and voilà, you have a bearing. You can also have a universal joint. This is again purely hydrocarbon, nice molecular structure.
Here is something that was done at NASA Ames a few years ago modifying buckytubes so you can have gears and stuff like that, some basic mechanical structures. If you go beyond hydrogen and carbon, you can start to talk about more complex structures. These are, how do I put it? If you go beyond hydrogen and carbon, then from a mechanical perspective you get a little increase in your set of capabilities but it’s not a vast increase.
Here, for example, we have a bearing. We could make a bearing of similar size out of hydrocarbon, maybe not quite as small, and it would be fine. If you have the additional elements available, you can go and do something a bit more flexible. Here is a planetary gear. This is something that would be familiar in a modern hybrid automobile; they have planetary gears inside them to play games with how they manipulate power and reassign it from the drive wheels to the batteries to the engine.
Planetary gears can be made very small. Again, this is a planetary gear that happens to involve more elements than hydrogen and carbon, but you could make a hydrocarbon planetary gear if you wanted to. It would probably a little bit larger. This particular planetary gear is something that people would be familiar with, except for the very small size. The other thing about using multiple elements is that you make them colorful. This is very important when you are talking with the press. The reason that you see this thing in the press coverage and not one of the hydrocarbon structures is because the hydrocarbon structures are black and white images. The press likes color, unless of course you are doing a black and white press run for a newspaper or something, in which case they’ll go for the hydrocarbon bearing.
Audience: There is the possibility of using functional color coding.
Yes, with hydrocarbons you could indeed show stress, as opposed to the color coding here, which is used to show which elements are involved. This planetary gear was simulated by some guys down at Caltech, Bill Goddard and his group, and it looks like it actually works. We had some interesting discussions. There was an earlier version that was driven very hard, up into the terahertz frequency, and it managed to skitter and skip. If you slow it down, it seems to work reasonably well. The overall lesson behind this is, I wouldn’t stake my life that this exact design is going to work, because you might have to tweak it a bit, but something pretty close to that design pretty obviously works. It is very clear that you can make small planetary gears on the order of this size. The conclusions you draw from this kind of design effort are that this kind of structure looks like it’s feasible.
Audience: How does that compare to the bacterial flagella motor?
The biological systems have… how do I put it? Their performance is much worse in terms of rotational speed, power, efficiency, you go down the line. What’s remarkable about the biological systems is not that they work well, but that they work at all. I mean, my gosh! You did what, out of what? In water? And it works? That’s really phenomenal.
Audience: Do the multi-element designs vibrate more?
Yes, if you look at this thing, it can vibrate. Certainly, if you look at multi-part designs, one of the things you have to be concerned about is, are they going to wiggle and vibrate? One of the things you do is make it so the barriers to rotation or the barriers to proper motion are very low so that it doesn’t have much opportunity to get into sympathetic vibrations at different frequencies. There have been simulations of various molecular bearings. If you run them fast enough, they start doing all sorts of weird things and breaking up, so the slower you run them then the more stable they are against undesired vibrational problems. The same thing happens in a motor.
If you’ve got a design where you’ve made your barriers very low and it rotates very smoothly, then you can slow it down and slow it down and not have to worry about speed. There are certainly designs in electronics where you have to operate at a certain speed and if you go too slow the charge dissipates. In mechanical designs I think there are things that would create similar problems, but we don’t go in for complicated designs. Simple designs that work semi-statically are the rule at this point in time because they are easier to analyze.
Robert Freitas: The gear can go up to a gigahertz.
Audience: What kind of simulation did you perform?
Ralph Merkle: The Caltech people did a molecular dynamic simulation. They started at zero speed and ramped it up with force and then began watching it do something useful and interesting. I think the first time they did it, they took an earlier design and twisted it with some incredibly large force and ran it at a terahertz. It didn’t work very well at that speed. It’s an interesting problem because obviously if you’re doing computational modeling you have a limit in the number of computer cycles you have, and one of the things you do is you want something interesting to happen, so you take your computer power, you figure out how fast you have to run the thing for something interesting to happen, like I want it to turn three revolutions or something, and then you run it at a speed which will give you that something interesting. Now, if it happens you have a small amount of computer power, then the speed is going to be very, very high, and you might well break the thing by operating it a lot faster than anyone was anticipating.
Audience: I heard a year or two ago there were things coming in from England from Phil Moriarty and others saying that you really can’t simulate some of these strange forces when you get down to the quantum level. What is missing in terms of the ultimate accuracy that you would like in computer modeling of something like this?
The answer is, if you’re looking at a structure like this, you have a specified set of bonds, it’s a stable structure, and you can model it using molecular mechanics. Molecular mechanics models the forces between the atoms, and it models the atoms themselves as Newtonian point masses. It turns out that there is something called the Born-Oppenheimer approximation, which is a fancy way of saying atoms are heavy, electrons are light… if I want to model the behavior of a system involving atoms, I can basically treat the electrons as a cloud that just settles into the ground state, given a fixed location of the nuclei, and therefore I can model the whole system quite accurately by saying the nuclei are point masses and the electrons provide the electronic structure, which describes the forces between those point masses.
Robert Freitas: The Born-Oppenheimer approximation is the quantum mechanics simulation, not a molecular mechanics simulation. That’s what’s missing. They’re arguing that you can’t simulate these things with complete certainty just using molecular mechanics, because possibly some of those gears that are hitting each other might form a chemical reaction, something like that. To check that out, you have to use a simulation package that can model that sort of thing, which would be density functional theory or something like that. Right now you can do density functional theory up to a few hundred atoms conveniently. The one that was up before is several thousand. We think there might be some packages where we could do DFT on several thousand atoms. But we have no reason to believe that it would fail, based on what we know.
This is a neon pump. This is a mechanical device. The core on the right rotates and is actually inserted into the main part of the pump. The idea is that neon atoms migrate along. You’ll notice at the core there is a helical structure in the middle, and as you rotate it’s been designed so that the neon atoms slide along and will come out the other end when you rotate it. Again, this has been simulated by Bill Goddard’s group down at Caltech. They looked at it and said, “Gee, it looks like it actually pumps neon.” It probably is not as selective as we might have liked. It probably will pump a few other small atoms in that size range, but it looks like it’s a pretty good neon pump.
Robert Freitas: That is also a rotary motor driven by gas pressure if you operate it in the opposite direction. Push gas through it and you’ll get rotary motion. It’s a really nice motor.
Ralph Merkle: Most of these structures are fairly stable. In other words, if you look at them people will say, “Oh yeah, that’s a fairly boring structure.” In fact, one of the problems we have, particularly when you go to hydrocarbon structures, is chemists look at it and say, “That’s really boring. You don’t want me to publish something about a boring structure do you? I want an interesting structure.” Where “interesing” means we wouldn’t want to touch it.
Robert Freitas: It’s on the verge of falling apart.
Ralph Merkle: It’s “interesting.”
Okay, a quick overview on making diamond today. One of the ways of doing it is with chemical vapor deposition. Chemical vapor deposition involves making a highly reactive gas, and that gas involves things like hydrogen and some form of carbon, so you have reactive species of carbon. You add energy in some form, so you wind up with very reactive species of carbon. CH3 for example is a reactive species of carbon. It’s got a dangling bond; it wants to react. You need atomic hydrogen. I won’t get into the reason you need atomic hydrogen, but typically in these plasmas you have a lot of atomic hydrogen. The result is that you react with the surface. So you abstract atoms from the surface. In particular if you have a hydrogenated diamond surface, then atomic hydrogen will react with the surface, pluck off a hydrogen, leaving a dangling bond on the surface. A reactive form of carbon can then react with the surface, and then you get the growth of diamond on the surface.
So that’s a fast introduction, a synthetic strategy for the synthesis of diamond. The synthetic strategy is really simple. You have tools. Tools can be moved around. If the tool can be moved with six degrees of freedom - you might be able to get away with fewer in various circumstances, but let’s just talk about six in the general case. The tool comes up, the tool is highly reactive, it reacts with the surface, you pull the tool away, the surface has been modified, you pick up another tool, it reacts with the surface, you pull the tool away, the surface has been further modified.
By using a sequence of successive modifications of the surface, you can then build up the structure you want, one step at a time, using molecular tools. If you are going to have highly reactive tools, they had better be in an inert environment or they will react with the environment. Vacuum is a really nice inert environment, so we can just assume we are doing things in vacuum.
Finally, what would be a molecular positional device? If you’re going to talk about molecular tools, what about a molecular positional device. Here is a proposal for a molecular positional device. It’s a robotic arm. There are other proposals floating around as well. The idea is that something like this would be on the order of one hundred nanometers tall and would allow you to manipulate structures and build them using positional synthesis and positional control.
Now we start to get into the stuff we have been doing more recently. There is an annotated bibliography on diamond mechanosynthesis, which has over fifty entries in it. It discusses a variety of different reactions that are involved in the mechanosynthesis of diamond.
How many have seen this before? This is a hydrogen abstraction tool. The idea here is that we are going to have the ethynyl radical, which has a higher affinity for hydrogen than almost anything else. You bring this ethynyl radical up to a diamond (111) surface. Here comes the hydrogen abstraction tool and it just plucked off a hydrogen. The hydrogen abstraction tool then pulls away from the surface. Isn’t that amazing? We selectively removed a single hydrogen atom at room temperature from a diamond (111) surface. That illustrates the basic concept of mechanosynthesis.
This is using something called Brenner’s potential, which is a molecular mechanics force field, which can handle the making and breaking of carbon-carbon and carbon-hydrogen bonds, and therefore is useful for this kind of simulation. The hydrogen abstraction tool has actually been fairly well studied. The claim is that in all talks you should have one slide which is either incomprehensible or just has too much stuff on it for anyone to follow.
So, this is the slide. The idea is, you are supposed to look at it and say, “Wow, there are a bunch of references on hydrogen abstraction tools.” And there are. A bunch of people have looked at it using various theoretical tools, and the conclusion is always the same: “Gosh, gee, it looks like the hydrogen abstraction tool will pluck off a hydrogen from a hydrogenated diamond surface.
Robert Freitas: And they are quantum chemistry simulations.
Ralph Merkle: What it boils down to is if you burn more computer time you can get better results, and computational chemists understand how to burn lots of computer time. Give them more computer time, a giant supercomputer cluster, and they can give you better answers.
Audience: And if you take a computational chemist off the street, you’ve done a good deed.
Ralph Merkle: That too, yes.
Audience: As a software engineer, I think of “abstraction” as something totally different. Why do they call it “abstraction” and not “extraction”?
Ralph Merkle: It’s a chemical term, and I don’t know. It’s one of those mysterious things you find in the literature. There is also hydrogen donation, which is the reverse of abstraction. I often call it a hydrogen plucker, but that’s not considered good chemistry.
Audience: “Abduction” means to take away.
Audience: Why don’t you call it “rendition”?
Ralph Merkle: I’m going to avoid getting into the Bush jokes. That would just take up the rest of the time.
Now we’re up to the point where we can selectively pluck hydrogens off surfaces. This means we can have surfaces that are reactive in a site specific fashion. The next thing that I am looking at here is a dimer placement tool. This is a tool that would allow you to position two carbon atoms on a surface. And we are showing two phases here: one where the dimer placement tool is touching the surface and one where it’s withdrawn. We don’t have the video of it, but it actually was analyzed using density functional theory, ab initio analysis, and molecular dynamics using ab initio quantum chemistry, which means that we are pretty confident it works. It used quite a bit of computer time.
Basically, it shows a germanium-based tool coming up to a (110) surface. In this particular case we decided we liked (110) surfaces because they’re groovy surfaces, so to speak. This is at 300 Kelvin, and it looks like it works. This was work done back at Zyvex. We used a computer cluster, worked on it, and published some papers, so that looks pretty good.
Robert Freitas: That work was about 100,000 CPU hours. In that study we looked at what happens if you bring the tooltip down at various angles, various tilts, various lateral errors relative to the spots that you are aiming at, and we ascertained exactly what are the limits of error in each of those directions so we know exactly what our viable operating envelope is for that tool on that surface. That is the sort of study you would have to do with all the tools for all the surfaces you’re going to do.
Ralph Merkle: There is a lot of work that needs to get done. It’s not as though there is any shortage of work to do.
Audience: How many atom radii of error do you have there?
Robert Freitas: On that one it is about 20 to 30 degrees of tilt, and .2 to .5 angstrom is general placement accuracy just as a rule of thumb. It varies a little bit from that, but that’s the general range. If you have half an angstrom you can probably do entry level reliable mechanosynthesis with that.
Ralph Merkle: Also it depends on which direction you’re going. There are specific issues there.
One of the things that first happens when you talk about the general goal of mechanosynthesis, we want to make almost anything. Making almost anything is complicated, because you have over a hundred elements in the periodic table. If you want to be able to build structures that use all of the elements in the periodic table, which obviously we are going to do one of these days, then you’re going to need over a hundred tools. If you’re looking at the various structures you can build with over a hundred elements, there are a lot of them. If you put a few atoms down, suddenly there is a combinatorial explosion. So you need a lot of analysis to figure out what is going on. The advantage of course is that if you are talking about building things using the full periodic table you get a lot of flexibility in your synthesis. You can have complex tools that deal with all kinds of complex nice structures and you can do all kinds of fun things. You can make chemists real happy, using sophisticated chemistry. And you can build almost any structure consistent with physical law, because you’ve got the full periodic table to play with. The problem is, we don’t want to do that. We don’t want high complexity.
It’s real hard to come up with a structure that you can’t build. It’s difficult. It would require a lot of analysis. No one has even approached that. No one has published a structure saying, “You know, I think you can’t make this, nyaaa!” No one’s done that yet. Someday, someone’s going to have to do it, but so far they haven’t.
Robert Freitas: A gallery of unbuildable parts.
Ralph Merkle: So far there has not been a lot of analysis on that particular problem.
Now, we get to the new, exciting stuff that we are doing on the cutting edge of the frontier of scientific research. Rob and I have a paper called “A Minimal Toolset for Positional Diamond Mechanosynthesis.” It’s going to be appearing pretty soon. We’re not exactly sure which month, but it’s going to be coming up in a couple months.
How many people don’t like patents? This is a preliminary patent application. The rule of thumb in patent applications is you swamp the guy, then he gives it to you. Usually what happens is they say, “No, no, no, I don’t accept it.” And if you have enough raw verbiage, you can send it back with big responses explaining why it is he’s wrong and you’re right.
Robert Freitas: They only have six hours for each patent, so you divide that by the pages. They have about one second per page.
Ralph Merkle: There are various interesting things going on in the patent system. One of them is that the patent clerks are totally overworked. We’re making our contribution to taking advantage of this system.
At any rate, the paper is going to be appearing, and the basic idea is we’re going to trim the number of elements involved. First off, if we’re going to build diamond, we are going to need hydrogen and carbon. When we started looking at it, we saw we needed something, because the reactions that we’re looking at, we couldn’t quite figure out how to get nice clean reactions for the synthesis of stuff without tossing in something. Initially we thought we would have to toss in a few elements. Rob said maybe we could get away with just germanium. And gee, it looks like we can get away with just germanium.
The proposal is: hydrogen, carbon, germanium. This puts a sharp limit on the combinatorial explosion, it puts a sharp limit on how many elements you have to think about, so life is a lot easier. Hydrogen and carbon let you build a whole bunch of stuff. There’s diamond, graphite, buckytubes, fullerenes, carbine rods, various organic compounds, there’s a bunch of stuff you can make with just hydrogen and carbon. The germanium we tossed in to provide just enough synthetic flexibility so we could actually make everything work out.
Right now the focus is on a proposal that is simple enough that we can analyze it, but flexible enough that it clearly opens up a huge range of possibilities. That’s what we’re doing. We went ahead and did some analysis. We got ourselves a nice computational chemistry program called Gaussian. It’s a standard computational chemistry program. One of the things going on is when you publish a paper, part of the credibility of your paper is “We used Gaussian.” Computational chemists know what Gaussian is. They say, “Okay, I don’t know who these guys are, but I know what Gaussian is and I know the parameters that they used.” That should be good. Part of the credibility is simply the computational chemistry package.
So, we went ahead and analyzed it. And it’s 1630 tooltip/workpiece structures, 65 reaction sequences, 328 reaction steps, 354 unique pathological side reactions. One problem is, you bring a tool up to a surface, there is what you want to have happen, and then there is all the stuff you don’t want to have happen. You have to analyze all the stuff you don’t want to have happen to make sure it won’t happen. We had to spend a lot of time analyzing stuff that we didn’t want to have happen to make sure it wouldn’t happen. Then we had to throw stuff out and start over again on various occasions.
We didn’t do molecular dynamics; what we did was do specific analysis of the energies of a reaction. Then we’d eyeball it and say, “You know, it’s conceivably possible that this hydrogen could go off and bond with that carbon, or this carbon could slip in the following ugly conformation.”
Robert Freitas: We spent a lot of time trying to postulate all the possible scenarios. One of the reactions, I think it had ten different pathologies associated with a particular reaction. Usually, it’s more like three or four. If the pathological states are preferred energetically to the one we want, it’s thrown out. If two are very close in energy, that means the reaction isn’t reliable enough. It might work half the time, but the other half it won’t. We throw that out. It’s got to be clearly energetically preferred by a level of energy margin that we set at the beginning of the paper that we believe gives us enough reliability that we can use that. All those 65 reaction sequences pass all the tests.
Ralph Merkle: What’s nice is that the set of reactions is complete.
Here are the tools. There are nine tools. Starting from the upper left, we have the hydrogen abstraction tool with the ethynyl radical. This is the tool that has been studied so extensively in the literature, and we are pretty comfortable with it. The next one is the hydrogen donation tool. You’ll notice it has a yellow atom. That yellow atom is a germanium. The germanium-hydrogen bond is a fairly weak bond, so this tool can be used to donate a hydrogen to a growing carbon structure. If you’ve got, for example, a carbon radical, a dangling bond on a growing structure, you can bring up the hydrogen donation tool, donate a hydrogen, pull the tool back, and the hydrogen stays on the surface.
We have the GM tool, which has got germanium and a methylene on top. The methylene is reactive. Again, the gemanium-carbon bond is supposed to be weak, so if you bring this tool up to a surface where there is a dangling bond, the reactive carbon on the tip of the tool will react with the surface. The germanium-carbon bond is weak, so this tool is useful for taking a carbon atom and plopping it onto a surface specific location.
Then we have the germylene tool. That’s sort of the flipside. That’s got a carbon at the bridgehead position and a germanium atom, so now you can use this tool to bring up a germanium to a structure. You have to go through more antics in order to pull back and make sure that your germanium actually sticks with the surface and not with the tool, but you can make that happen.
We’ve got a methylene tool. You can make it work if you’re patient with it. It’s got a carbon-carbon bond, so when you bring it up to a surface obviously you’ve got a problem in the sense that if you’re bonding a carbon to a surface and you’ve also bonded a carbon atom to the tip of the tool, then basically you’ve got a tug of war, you have to play games in order to make it come off in the right way. It’s more complicated to use, and we’ve got some sample reaction sequences.
The hydrogen transaction tool is simply the germanium radical brought up and bonded to the hydrogen abstraction tool with a hydrogen on the tip. It turns out that if you do this, the hydrogen on the tip is now very weakly bonded to the carbon atoms. The hydrogen tip is now going to fall off at any excuse. So this is a great way to donate a hydrogen to a structure. If you pull back on the germanium, the germanium will then leave that structure, and leave the hydrogen behind. You can use that particular tool both to weaken the bond and also to strengthen the bond.
Then we have a simple adamantane radical, which you can use as a carbon radical. We have the dimer placement tool, which we discussed earlier. That’s got two germanium atoms and uses a dimer to place that on a surface. And finally, a germanium radical. That’s it. Nine tools. Period, end of discussion. We have complete recharge sequences for all the tools.
Robert Freitas: I guess it goes without saying, these are all built on an adamantane base, which gives you the assumption that you can attach it to a larger structure, which is a regular diamond lattice.
Audience: How much do Vanderwahl forces come into play?
Ralph Merkle: The answer is it’s there, but it doesn’t prevent it from working. The explicit simulation was using small structures to model the chemical reactions at the tip.
Robert Freitas: We have proposals in the paper for how to build three of these tools using only current technology. We have an experimentalist who has told us that he thinks it looks reasonable, and he wants to try doing it. That’s as far as we’ve gotten with that.
Ralph Merkle: The purpose of this was to get a complete set of reactions, because this has not been done before: enumerating a set of tools and enumerating the full set of reactions involved in making those tools from raw material, and also showing that you can build a wide range of interesting other structures. It just hadn’t been done.
Audience: A lot of these things predicate that you have a diamond substrate surface. With current diamond deposition technology, how big a patch could you make?
Robert Freitas: It’s very easy to start building these structures if you start with an adamantane cage. You can build more cages onto the first cage. So you really only need a tip which has a cage at the bottom of it, and you can use that to build off of. You don’t need to start with a flat surface.
Ralph Merkle: If you have a set of tools, the question is how do you manipulate them? You could start by using some descendant of existing scanning probe microscopes. You can position something using a macroscale device. These could also be used at the tip of a molecular scale device like that robotic arm I was showing. There are ways you could control that remotely. If you’re talking about molecular positional devices, first we have to get the molecular positional devices and then we have to discuss how it is you control those remotely. That’s another step along the path. We’re currently working on the proposal for what is the grand scheme for putting all of this stuff together.
I should mention, by the way, that the reason we went ahead and did this work is that when we were at Zyvex, we went through a complete design for an assembler, discussed it with everybody, discussed what things they were concerned about, and the general concern there, as well as other places where we talked about it, was exactly what are the chemical reactions? If you have a complete analysis of the specific chemical reactions, this puts to rest a whole set of pressing concerns. I think we’ve run out of time at this point.
posted at: 23:33 | path: /sci/nano | permanent link to this entry
