I’m working on a series of posts that will describe practical research objectives on paths toward advanced nanosystems. My main concern is progress in fabrication, and “Modular Molecular Composite Nanosystems” can be seen as first in the series.
This post, however, will be atypical. It’s about impractical research objectives that have received far too much attention. I want to get this topic out of the way at the beginning. Some widespread ideas about research objectives
- are bad
- seem absurd to most scientists
- are inconsistent with my ideas and publications
- are nonetheless widely attributed to me
I really dislike ideas like these, and all the more so when the ideas have spawned a jumble of misconceptions that impede progress. Some ideas about diamond synthesis are are in this category.
Why diamond synthesis is a bad objective
Not now, please
Contrary to popular opinion, diamond synthesis seems almost irrelevant to progress toward advanced nanosystems. At the current stage or research, it is both difficult and unnecessary. In a following post I’ll present some criteria and metrics for judging materials, praise cerium dioxide, and explain why humble fool’s gold is better than diamond.
At the outset, I should note that diamond is in many ways a true wonder-material: It simultaneously excels in hardness, strength, rigidity, thermal conductivity, electron mobility, and several other properties prized by engineers. This is why applications for diamond keep growing.
Diamond does, however, have a grave shortcoming: synthesis has been, and continues to be, difficult and expensive. Stubbornly so. The methods I’m familiar with require either high temperatures and ultrahigh pressures, or highly reactive gas-phase species interacting with a hot surface in a vacuum chamber. Neither process is suited to atomically precise control. Advanced mechanosynthetic methods (of the sort analyzed in Chapter 8 of Nanosystems) will eventually erase this problem, but the emphasis here is on the words “advanced” and “eventually”.
Diamond is a great material, but it’s hard to make, and there are other materials in the world that serve well in demanding tasks. In light of this, the intense focus on diamond nanotechnologies seems quite mysterious: How did the idea of molecular manufacturing — a general approach to organizing mechanosynthesis — been become so closely identified with making diamond? And consider “mechanosynthesis” itself. The term means molecular synthesis directed by mechanical means, nothing more, and this is a concept broad enough to embrace the synthesis of proteins by ribosomes. How did such a basic and generic concept become equated with mechanosynthesis of diamond?
And most important: Considering the difficulties of diamond synthesis, why treat diamond mechanosynthesis as if it were a necessary first step toward molecular manufacturing? Building a tiny bit of diamond this way would of course be an impressive lab demo, but the plausible technologies for achieving this seem difficult to extend, and I doubt that they would be very useful in any general sense.
Indeed, a leading nanotechnology company, Zyvex, used computational modeling (density-functional based molecular dynamics) to study diamond mechanosynthesis and concluded that, although the physics would work, but the present-day laboratory practicalities would not. They’ve chosen instead to pursue mechanosynthesis of silicon-surface structures, using a clever approach termed “patterned atomic-layer epitaxy”, a method in which the only mechanically directed operation is the removal of hydrogen atoms from a hydrogen-passivated surface.
About the confusion
Nanosystems discusses diamond and diamond-related materials at length, and the ideology of diamondism grew from my own words. It is, however, a mutant, alien growth. From my first paper forward (1981, 1986, 1992, 1994, 1999, 2005,…), I’ve advocated developing biomolecular nanosystems and building from there, and this is what’s been happening.
In Nanosystems, I discuss diamond synthesis not an early step in development, but as a particularly difficult test-case for the application of advanced mechanosynthesis to high-performance materials. The confusion about this has been so annoying that I’ve posted the relevant paragraphs from Nanosystems here, just to set the record straight.
In short, there’s is a huge difference between a practical, near-term objective and an attractive but distant aim point. It’s time to shift attention to aim points that are much closer. I’ll write soon about materials, criteria, metrics, and more appropriate materials. (Aqueous synthesis at room temperature, for example, is very nice.)
Title updated 10 Feb 09
Related posts:
- Modular Molecular Composite Nanosystems
Biomolecular engineering for atomically precise nanosystems - Toward Advanced Nanotechnology: Nanomaterials (1)
Why I’ve never advocated starting with diamond - Toward Advanced Nanotechnology: Nanomaterials (2)
Stiffness matters (and protein isn’t remotely like meat) - Self-Assembly for Nanotechnology
The virtues of self-assembly and the benefits of external guidance - From Self-Assembly to Mechanosynthesis
Mechanosynthesis begins with soft machines - Toward Advanced Nanotechnology: Nanomaterials (3)
Mechanical engineering meets thermal fluctuations - Toward Advanced Nanotechnology: Nanomaterials (4)
Nanostructures, Nanomaterials, and Lattice-Scaled Stiffness - Toward Advanced Nanotechnology: Nanomaterials (5)
Nanomachines, Nanomaterials, and Klm
{ 4 trackbacks }
{ 14 comments }
Esteban 12.27.08 at 9:51 am UTC
I think that it must build other thing before that diamonds. It must
build atomic transistor to create supercomputers of xeraflops, this will accelerate dramatic the science and the technology. I think this will
achieve very soon( 2-3 years ), groups of Spain and Japan are doing
progress to manipulate atoms with AFM.
Eric Drexler 12.27.08 at 7:06 pm UTC
Using self-assembly or guided molecular assembly to build something as complex as a computer is an advanced objective, several big steps beyond current capabilities. Electronic devices based on single atoms (but much bigger overall!) have been made, but to use them in a computer, the devices would have to be made reliably and integrated with nanoscale lithography. This will take more than a few years. The 2007 International Technology Roadmap for Semiconductors has a section on Emerging Research Devices [pdf], and this includes an evaluation of the practical potential of molecular memory and logic devices. The evaluation is much more positive for memory than for logic, but no specific technology is identified as a likely prospect.
I think there’s a good chance that practical self-assembly approaches can be developed to implement multi-layer, high-density non-volatile RAM (on a lithographic substrate), because the necessary structures would have crystalline regularity, and because memory systems of this sort can be tolerant of defects in the self-assembled part. A capacity on the rough order of 1015 bits in a chip-size package seems achievable.
Longer term, directed assembly methods will enable the implementation of intricate structures, new devices, and massively parallel computational systems of the sort you have in mind. A conservative physics-based analysis indicates that a billion-CPU laptop computer is well within the bounds of what is possible.
kurt9 12.27.08 at 8:15 pm UTC
Hear, hear!
Back “in the day”, I and several others had this discussion with a certain Thomas Donaldson (remember him?). We all agreed that nanotech, once fully developed, is likely to be “wet”, using processes similar to those that go on in biological systems. This is a position that I still believe today and seems to be the direction that the research is going.
The issue of “wet” or “dry” nanotech is a red herring. The real issue is to develop any kind of automated exponential manufacturing that will allows us to make the things that we want without having to put up with all of complaining and whining of those we want nothing to do with.
At least that’s how I look at it.
Esteban 12.28.08 at 9:35 am UTC
Prof. Drexler, I share with you that commercial level will not have atomic computer in 2-3 years, but experimental is possible. Definitely
this year Dr. Anirban Bandyopadhyay( Japan MNI ) has made molecular processor with 16 duroquinone molecules. On December
of the 2009 will have made a 1024 molecules processor ( with a new software didn`t base in differencial equations, Stephen Wolfram )
I think this processor will evolve to atom processor.
Also I think this next year 2009 will be a year of new paradigms,
for example the use of petaflops computers ( Jaguar,Roadrunner,etc )
to do atoms simulations, it will much accelerate these projects.
Excessive optimism?
Happy New Year 2009 !!!
alexandro 12.28.08 at 2:44 pm UTC
what about nanomedicine? aren’t Freitas describing diamond -based solutions in his work?
Perry E. Metzger 12.29.08 at 4:32 pm UTC
With respect to a far greater mind than mine, I think you’re wrong, Dr. Drexler.
The situation reminds me of John McCarthy’s original reluctance to believe the simplest way to build a lisp interpreter was to hand compile the “eval” function — McCarthy insisted that “eval”, as defined in his published paper, was simply an academic proof of concept and not intended to be a blueprint for a working system. Luckily for the world, his student ignored him. Equally luckily, McCarthy admitted he was wrong once he was presented with the evidence.
So far as I can tell, as enormous as the barriers to research on direct to diamondoid are, the challenges are not particularly larger than those in attempting to bootstrap molecular manufacturing from protein engineering or structural DNA technologies. Furthermore, there are actually substantial advantages to the path: the fact that diamondoid structures are rigid makes them far more amenable to engineering analysis, and the fact that going by the other paths means that it will be necessary to create a whole technology that then gets thrown away to produce more advanced technologies.
Although five years ago I would have agreed with you (and probably would not have had the personal confidence to disagree with someone as eminent as yourself), I no longer do. I’ve been looking extensively at the issues involved, and although they’re substantial, I think the principle obstacle right now to significant progress along the diamondoid path is a lack of money and researchers. I am not sure that obstacle can be overcome, but it is not per se a technical obstacle.
jim moore 12.29.08 at 6:40 pm UTC
“Building a tiny bit of diamond this way would of course be an impressive lab demo, but the plausible technologies for achieving this seem difficult to extend, and I doubt that they would be very useful in any general sense.”
Could you elaborate on that statement a little bit. Are you saying that even if Dr. Moriarty’s group is successful in experimentally making and using the Freitas / Merkel dimer placement tool tip to add a pair of carbon atoms to a precise spot on a diamond surface it would not likely lead to a useful fabrication technology? If so, could you explain why it would be unlikely to lead to useful fabrication technology.
Phil Duby 12.29.08 at 9:56 pm UTC
Can’t FPGA be used to build a computer (logic circuits), and shouldn’t FPGA have crystalline regularity? More complex than memory, but still a basic repeating pattern.
Erin 12.30.08 at 4:11 am UTC
First of all thank you for this excellent blog, Dr Drexler.
I have been absorbing information on nanotechnology for years now and I think that diamondoid and similiar materials are excellent for mechanosynthesis, but I also think other materials may be more ammendable to nearer term fabrication, such as silicates and polymers. There were some excellent papers written by this man Stephen Gillette about that topic, I am sure you’re aware of them. He says that silicates can be polymerized from liquid phase and at more ambient temperatures, and are also hard and strong and can be used as a bridge to diamondoid materials. One great thing, materials like silicon and aluminum oxides dont burn because they are already oxidized, whereas even fullerenes and diamond can burn at high temperatures in oxygen.
What about metals? In Nanosystems you said only the hardest stiffest strongest metals with the highest melting points are worth it for MNT, I tend to agree.
But is there room for improvement with metal alloys and nanotechnology?
Markku Jantunen 12.30.08 at 4:13 pm UTC
The 2007 International Technology Roadmap for Semiconductors has a section on Emerging Research Devices [pdf], and this includes an evaluation of the practical potential of molecular memory and logic devices. The evaluation is much more positive for memory than for logic, but no specific technology is identified as a likely prospect.
Space and time can be traded in algorithmics to varying degrees. If it should be possible to use molecular manufacturing to mass-produce cheap memory circuits of stupendlously large capacity (and fast access/write speeds), wouldn’t it make sense to use them to build computers (at least for special purposes) whose processors are relatively inefficient but whose memories are extraordinarily large, particularly if on chip photonic communication between components becomes commercially available?
Eric Drexler 12.30.08 at 11:19 pm UTC
@ Erin — Yes, I know Steve and agree that silicates could be very useful in building a range of atomically precise structures and devices. I expect that they will be. Some silicates have a special advantage for intermediate-stage developments: They can crystallize spontaneously from aqueous solutions. This property suggests that biomolecular self-assembled structures (and productive nanosystems) could shape closely related structures by guided growth processes. (Some other oxides share this advantage, however, and have better mechanical properties — this is important for some but not all uses).
@ jim moore — You asked me to elaborate on what my remark that building a tiny bit of diamond would be an impressive lab demo, but that I don’t think its likely that the methods could be extended to produce results that are very useful in a general sense. What I have in mind is the relative difficulty and rewards of competing approaches, some of which have gotten far too little attention. I’ll be saying more about this, of course.
When looking down the road, it’s important to distinguish current, intermediate and advanced technologies and capabilities. Diamond and related materials are excellent, and that using them in nanostructures will eventually be practical for even large-scale applications. My remarks are directed to current methods, extensions of them, and intermediate developments
@ Markku Jantunen — Yes, larger memories shouldn’t be regarded as merely providing more space for existing software, but as an opportunity to restructure software to shift the memory/processor trade-off. A technique that I’ve used from time to time is memoization.
@ Phil Duby — As you say, FPGAs can be implemented with highly regular structures, and this makes them candidates for relatively early development. They’re a way to make complicated circuits from simple structures while avoiding complex fabrication. In comparison to memory, there would be a greater premium on speed, and the unit cells would of course be more complex. Ensuring tolerance for fabrication defects would also be more complex, but an effective approach is to locate the defects and then customize the mapping of the logic circuits to the physical structure, as demonstrated in the Teramac work [pdf] at HP. Their machine worked despite a 10% defect rate in its programmable logic cells.
Erin 12.31.08 at 4:51 am UTC
Doctor Drexler, thank you for your response. I have been inspired by you and your work in nanotechnology over the years. And this leads to a question I have: I graduated from high school but apart from that I did not pursue further education, apart from study through reading on my own time. I am seriously considering going back to school, I want to get involved in nanoengineering, but what is best for me to do? I was told I should go and start with a two year degree in chemistry. What are your thoughts, Drexler? What would be the best way for me to study something that will allow me to contribute to actual hard developments in MNT?
Perry E. Metzger 01.01.09 at 3:18 pm UTC
Erin: I would suggest getting a four year degree in chemistry, you just aren’t going to get enough out of a two year degree program unless you’re doing it purely as an experiment to see if you are comfortable with school. You want a solid grounding in organic and physical chemistry — the organic to understand what complicated reaction mechanisms look like, the p-chem to understand the underlying physics behind those reactions and molecular systems.
Although I’m not sure molecular biology is a must, having a class in it is not a bad thing, if only because biology is the only worked example we have of a fully functioning nanotechnological system (though one would hope that engineered systems would be far less messy and disorganized.)
Many budding chemists skimp on learning physics — don’t do that. Physics is pretty important too.
Matt 01.05.09 at 1:41 am UTC
Kurt9:
I agree with your red herring comment. Building automated directed assembly systems will be more difficult than wet or dry mechanochemistry. At Zyvex we called this the “systems design” problem but I’m not sure we made significant progress toward any solutions. Some combination of directed assembly and self assembly will be necessary. I don’t have much hope for scanning probe based techniques as a manufacturing technology. I’m sure some great R&D will continue to be done with these tools - but how these tools can make significant quantities of anything is a mystery to me.
Comments on this entry are closed.