A couple of journalists have recently asked me some questions about the EPSRC Ideas Factory on software control of matter that I am directing in January. The obvious question is whether software control of matter – which was defined as “a device or scheme that can arrange atoms or molecules according to an arbitrary, user-defined blueprint” – will be possible. I don’t know the answer to this – in some very limited sense (for example, the self-assembly of nanostructures based on DNA molecules with specified sequences) it is possible now, but whether these very tentative steps can be fully generalised is not yet clear (and if it was clear, then there would be no point in having the Ideas Factory). More interesting, perhaps is the question of what one would do with such a technology if one had it. Would it lead to, for example, the full MNT vision of Drexler, with personal nanofactories based on the principles of mechanical engineering executed with truly atomic precision?
I don’t think so. I’ve written before of the difficulties that this project would face, and I don’t want to repeat that argument here. Instead, I want to argue that this mechanically focused vision of nanotechnology actually misses the biggest opportunity that this level of control over matter would offer – the possibility of precisely controlling the interactions between electrons and light within matter. The key idea here is that of the “metamaterial”, but the potential goes much further than simply designing materials: instead, the prize is the complete erosion of the distinction we have now between a “material” and a “device”.
A “metamaterial” is the name given to a nanoscale arrangement of atoms that gives rise to new electronic,magnetic or optical properties that would not be obtainable in a single, homogenous material. It’s been known for some time, for example, that structures of alternating layers of different semiconductors can behave, as far as an electron is concerned, as a new material with entirely new semiconducting properties. The confinement of electrons in “quantum dots” – nanoscale particles of semiconductors – profoundly changes the quantum states allowed to an electron, and clever combinations of quantum dots and layered structures yield novel lasers now, and the promise of quantum information processing devices in the future. For light, the natural gemstone opal – formed by the self-assembly of spherical particles in ordered arrays – offers a prototype for metamaterials that interact with light in interesting and useful ways. This field has been recently energised by the theoretical work of John Pendry, , at Imperial College, who has demonstrated that in principle arrays of patterned dielectrics and conductors can behave as materials with a negative refractive index.
This notion of optical metamaterials has achieved media notoriety as a route to making “invisibility cloaks” (see this review in Science for a more sober assessment). But the importance of these materials is much more general than that – in principle, if one can arrange the components of the metamaterial with nanoscale precision to some pattern that one calculates, one can guide light to go pretty much anywhere. If you combine this with the ability from semiconductor nanotechnology to manipulate electronic states, and from magnetic nanotechnology to manipulate electron spin, one has the potential for an integrated information technology of huge power. This will probably use not just the charge of the electron, as is done now, but its spin (spintronics) and/or its quantum state (quantum computing). There are, of course, some big ifs here, and I’m far from being confident that the required degree of generality, precision and control is possible. But I am sure that if something like a “matter compiler” is possible, it is manipulating photons and electrons, rather than carrying out fundamentally mechanical operations, that its products will be used for.
Richard, to be fair, one of the long-term goals of the Drexlerian project is to use linked nanodevices that can be made to mimic different materials, possibly even, on-the-fly. This would be a necessity if one is to turn diamond into a material that can be used for any application. Without the ability to use diamond to make such smart-materials, you would have to expand the number of reaction types included in a nanofactory to use many more elements in the periodic table. That becomes problematic for temperatures reasons that you are quite aware of, i.e. your six challenges.
On the issue of quantum dots and wells, the most ambitious treatment of these ideas I’ve seen is Will McCarthy’s “Hacking Matter.” I have not read the book, but rather an article about the idea in Wired magazine. The main concept being something called “wellstone,” a material composed of programmable quantum wells that can change in an instant. When I first heard of these ideas I thought they were, if anything, more outlandish than Drexlerian MNT. Upon latter reading, it seems that a lot of progress in this area has happened. Whether this leads to something as out-there as McCarthy’s “wellstone,” who can tell?
An interesting question arises, how would a matter compiler differ if you used quantum dots and wells not as finished products, but as part of the manufacturing process itself? The comparison has been made between the current designs of nanofactories to the computer designs of Charles Babbage. Babbage’s mechanical approach would have worked. In 1991 a working difference engine was built settling the debate of the design’s feasibility. In the end, however, we know there never was a true golden age of mechanical computers. By the time the demand for computers grew enough, the mechanical gave way to the electrical.
I have often wondered how the design of a nanofactory would change if you took in all the advantages of quantum processes. For example, one of the key components in any nanofactory design is the sorting rotors needed to purify the feedstock. I have seen examples of using lasers to sort atoms, might a quantum dot laser prove to be more useful than rotors? Also, when it comes to guiding reactions would it be possible to have, as a tip, a small piece of programmable mater that you could change on-the-fly to make it release a molecule, thereby giving you a new solution to the so-called “sticky finger” problem? And finally, there is the problem in any bottom-up manufacturing process of quality control. In living systems if a part is accidentally malformed, it is simply broken down and reused. In a human sized factory if a part doesn’t come out right, you can simply look at it to tell if it needs to be replaced or discarded. The question is, in man-made bottom-up manufacturing how do you deal with errors? For chemists, if you destructively observe a molecule it doesn’t matter, there are plenty more where that came from. In a nanofactory environment it’s been proposed that by building one layer at a time, you can check your work non-destructively by an SPM tip. Quantum mechanics may allow another possibility via interaction-free measurement. This is the same idea used in the quantum bomb detector thought experiment, where you can get information about a system by splitting a single photon into two paths, one where the interaction occurs and you get information at the expense of loosing a bomb, and another where nothing happens. It doesn’t matter which universe, if you subscribe to the many-worlds interpretation, you live in. You get the result regardless. By using the quantum-Zeno effect you can improve the probabilities of success arbitrarily, 70%, 85%, etc. This idea was used in a quantum computing experiment where the computer gave an answer without “actually running.” It might be possible to even use this to watch a Bose-Einstein condensate without destroying it. I wonder if there is a fundamental quantum mechanical way of *making* thinks, not just measurement or computation.
I know that you place greater importance on capabilities of finished products over manufacturing methodology, but it may be impossible to get everything you want in a finished product without having access to new ways of making things. Some of those new ways of making thinks might even look identical to proposals from the MNT camp, others radically different.
Here are some relevant links.
The complete book “Hacking Matter” available for free from the author’s website:
http://www.wilmccarthy.com/HackingMatterMultimediaEdition.pdf
The wired article on “Hacking Matter”:
http://www.wired.com/wired/archive/9.10/atoms.html
Quantum Seeing in the Dark
Paul Kwait, Harald Weinfurter, Anton Zeilinger Scientific American Nov 96
Overview of the quantum bomb detector and quantum Zeno effect:
http://www.dhushara.com/book/quantcos/qnonloc/qnonloc.htm
Paul Kwait on quantum computation.
Where he explains what it means to get an answer from a computer that “doesn’t run”:
http://cosmicvariance.com/2006/02/28/paul-kwiat-on-quantum-computation/
Richard, I’ve left a large comment with mutiple links that is sure to run a foul of you spam-filter. Please fish it out.
Ah hah. Despite your best efforts Richard, your name has come to be linked with MNT. You are even trying to force a holography linkage. Richard Jones the MNT proponent ahhaha.
Phillip, try not to gloat – it’s not becoming!
Nanoenthusiast, I’m aware of the Hacking Matter book, but I haven’t read it. I assumed that it took the very exciting work on semiconductor superlattices and heterojunctions at the time and extrapolated it. How did the bit about programming the quantum wells work?
I take your point about this already being part of the Engines of Creation vision. Actually, it is very explicitly signposted by Richard Feynman in his lecture:
“What could we do with layered structures with just the right layers? What would the properties of materials be if we could really arrange the atoms the way we want them? They would be very interesting to investigate theoretically. I can’t see exactly what would happen, but I can hardly doubt that when we have some control of the arrangement of things on a small scale we will get an enormously greater range of possible properties that substances can have, and of different things that we can do.
Consider, for example, a piece of material in which we make little coils and condensers (or their solid state analogs) 1,000 or 10,000 angstroms in a circuit, one right next to the other, over a large area, with little antennas sticking out at the other end—a whole series of circuits. Is it possible, for example, to emit light from a whole set of antennas, like we emit radio waves from an organized set of antennas to beam the radio programs to Europe? The same thing would be to beam the light out in a definite direction with very high intensity. (Perhaps such a beam is not very useful technically or economically.) …
“When we get to the very, very small world—say circuits of seven atoms—we have a lot of new things that would happen that represent completely new opportunities for design. Atoms on a small scale behave like nothing on a large scale, for they satisfy the laws of quantum mechanics. So, as we go down and fiddle around with the atoms down there, we are working with different laws, and we can expect to do different things. We can manufacture in different ways. We can use, not just circuits, but some system involving the quantized energy levels, or the interactions of quantized spins, etc.”
I guess the point I’m trying to make is that in Nanosystems there is not (as far as I can remember) any real discussion of quantum mechanics and how you can exploit it if you do have nanoscale control – it’s all “the principles of mechanical engineering applied to chemistry”.
I guess you should measure how profound an effect a manufacturing revolutions is by the utility of the novel products generated. This is difficult to do if some/all of the products have yet to be invented.
For diamond MNT, there is already an assumption an actuator of sorts is a viable product; an actuator is absolutely necessary for the tooltip’s functioning (and it isn’t a certainty). Precision diamond products interfaced with actuators would transform every existing industry on Earth radically. Diamond MNT would be analogous to the invention of fire or writing.
I’m not even sure what the above “metamaterials” category fulling encompasses. It seems to be an overlapping mishmash of different nanoscale properties. As far as I’m concerned the optics revolution has already happened with the invention of the laser. Magnet innovations are nice, but it is a very broad field of often non-interrelated engineering. Are new Rare Earth magnets for wind turbines a Metamaterial? New polymer magnets? Is it just magnetically-aided Q-dots? Is there anything else, maybe novel magnetic computer memory designs or some sort of magnetic visual displays?
I see some Q-dots as being useful for new types of solar cells, and that may be tranformative for power generation. Q-dots lasers make better consumer electronics technologies, but we are already in the midst of this revolution and it hardly seems tranformative. I see Metamaterials as being transformative for media and communications. Of course quantum computers will be huge and they will be enabled by fundamental nanoscale materials sciences advances.
Basically, I think the new term “Metamaterials” is just another word for transducers. And the key Q-dot technology is already interdisciplinary enough without muddying the waters further by tying it to other product technologies enabled by similiar materials science advances. Even with mature quantum computers, I can’t see all classes of metamaterials being as tranformative as Diamond MNT for the simple reason that Diamond MNT would enable cheap robotics automation.
Don’t get me wrong: I like Metamaterials. I think my GameCube has a blue laser I can use to play NHL 2005. I’m sure next generation metamaterials will enable a virtual reality gaming platform. But the assured Metamaterials revolution won’t be nearly as tranformative as a potential MNT revolution would be (if MNT works).
Phillip, I’ve clearly failed to explain what I mean here clearly enough. A semiconductor heterostructure is a metamaterial in the sense that as an electron traverses its nanoscale structure, it behaves as if it were moving through some “virtual material” with quite different properties. Sometimes these properties cannot be obtained at all in a normal homogenous material. An example we have now of this is a photonic band gap material, which has a very distinctive dispersion relationship such that in some frequencies a wave just can’t propagate. The new (so far, in the optical range, entirely theoretical) metamaterials proposed by Pendry is even odder in the sense that it has a negative refractive index. So far, what we’re talking about here are materials that are structured on a nanoscale, but which have long ranged periodicity, so their properties have a certain homogeneity. But, to give an example from the optical bandgap materials, if you locally take out some of the repeating elements (essentially making deliberate defects) you can localise photons (making cavity modes) or confine them to certain paths (like waveguides). With electronic metamaterials, you can do the same tricks with electrons. And if you combine an optical metamaterial with an electronic metamaterial (for example, putting quantum dots in an optical cavity made from two distributed Bragg reflectors, as illustrated here, then you can get subtle interactions between light and electrons of the kind that people are currently talking about using for quantum computing. The interest in using magnetic metamaterials isn’t to make new bulk magnets; magnetism is about manipulating the interaction between electrons via their spins. The magnetic multilayers used in GMR hard drive read heads are an example of this, but the prize here is this idea of spintronics, in which you do computing using the spin of the electron as the variable. All the examples I’ve given of metamaterials in current use (except the photonic band-gap ones) are essentially made by one-dimensional production of layers plus planar techniques like selective etching, combined with some crude self-assembly in the layers. But the example of making waveguides in photonic band-gap materials suggests that if you are able to control the structure of these materials in all three dimensions you will have computing/information processing devices that are triply more powerful than what we have now, firstly because they fully use all three dimensions, rather than the basically 2-d geometry imposed by our planar processing techniques, secondly because they can use the full potential advantages of light for transmitting information and, thirdly the huge extra power obtained by using not just the charge of the electron to store and process information, but also its spin and quantum state.
I’ve talked about using metamaterials to control electrons and light and their interactions; the second part of Nanoenthusiast’s comment essentially asks the question – couldn’t we use these to control atoms as well? Thinking about this, I think the answer is yes, in principle, as we certainly can confine atoms in optical traps, so if we really do have complete control of the optical fields, then we could thus control atoms as well. I can’t even begin to think how you might implement or exploit this!
I read a science fiction book recently, can’t remember which one, which had a very advanced and “magical” form of nanotechnology. Arbitrary molecular structures could be created, but they weren’t done with clumsy Drexlerian robot arms and mills. Instead, they somehow set up what was described as a 3D hologram, but using atoms rather than light waves. This would then cause the atoms to automatically position themselves according to the wave patterns of the hologram and presto, the molecular structure was formed. I don’t know if anything like this could work in reality but it sounds vaguely like the kind of thing being discussed here.
Hal, the book is Accelerando. I wondered about that part too. Excerpt copy/paste from page 50 of the free pdf version:
“The 3D printer is cranking up. It hisses slightly, dissipating heat from the hard vacuum chamber in its supercooled workspace. Deep in its guts it creates coherent atom beams, from a bunch of Bose–Einstein condensates hovering on the edge of absolute zero. By superimposing interference patterns on them, it generates an atomic hologram, building a perfect replica of some original artifact, right down to the atomic level – there are no clunky moving nanotechnology parts to break or overheat or mutate. Something is going to come out of the printer in half an hour, something cloned off its original right down to the individual quantum states of its component atomic nuclei. The cat, seemingly oblivious, shuffles closer to the warm air exhaust ducts.”
Page 55:
“Sleep cycles pass; the borrowed 3D printer on Object Barney’s surface spews bitmaps of atoms in quantum lockstep at its rendering platform, building up the control circuitry and skeletons of new printers (There are no clunky nanoassemblers here, no robots the size of viruses busily sorting molecules into piles – just the bizarre quantized magic of atomic holography, modulated Bose–Einstein condensates collapsing into strange, lacy, supercold machinery.) Electricity surges through the cable loops as they slice through Jupiter’s magnetosphere, slowly converting the rock’s momentum into power. Small robots grovel in the orange dirt, scooping up raw material to feed to the fractionating oven. Amber’s garden of machinery flourishes slowly, unpacking itself according to a schema designed by preteens at an industrial school in Poland, with barely any need for human guidance.”
When I first read this I immediately dismissed it. This same author, in a different book, used quantum entanglement to send information FTL which is a big no no. Recently, however, I was flipping through the Hacking Matter ebook and came across a basic description of the same idea. The idea is to use “atom lasers” derived from Bose-Einstein condensates to make a *real* “atom hologram”, as opposed to one make by a photon laser. Apparently, there at least two scientist who think something like this will work Lute Maleki of JPL and Pierre Meystre of University of Arizona. Will McCarthy is dubious about this, but believes some version of this might be used to build-up nanoelectronics one layer at time.
I wonder if any of these more exotic ideas with regards to “metamaterials” would need a major boost in instrumentation to examine them without disturbing them. If so, I think perhaps the only way is through the counterfactual, or interaction-free measurement idea. There may be phenomena of use to nanotechnologist that can’t be observed any other way. Only with a better grasp of what’s going on with such systems will we be able to tell what can and cannot be done.
Richard, if I could characterize your post here in one sentence, would it be accurate to say that you are saying different types of nanoscale transducers will interact in novel ways as research advances in the years ahead?
Hhhmm, that’ll be CHarles Stross then, an SF writer whose novels that I hve read so far i rather enjoyed, although I didnt like one of his short stories I read a few years ago.
Phillip, yes, different types of nanoscale transducers, but also different types of nanoscale logic devices.
Richard. This could be something that we mostly agree upon.
The interaction of new physics like metamaterials where greater control of matter provides greater control of light and magnetism is part of feedback cycle. The greater control of light and magnetism contribute to even greater control of matter and information.
Plasmons bridging photons to electronics Plasmons could enable optical computers that are 100,000 times faster.
Hypersound and acoustic lasers along with terahertz radiation are examples of growing control old forms of energy. The precise control of sound with physical structures is like the control metamaterials provide with light and radiation.
Metamaterials provide more control of light and radiation
Anything that gives us faster computers gives us better nanoscale simulations and other capabilities that drive our ability to understand and control things.
Anything that gives us more precise sensors and microscopes give us better understanding and control of what is happening at smaller physical scales or shorter time scales. We can see what is currently too small or too fast.
I still believe that the precise mechanical control of matter is a key contributor to this feedback cycle advancing our capabilities to handle more information, light, energy, matter and magnetism.
Recent announcements:
Advances in DNA nanotechnology. Nanorobotic arm arrays by Ned Seeman and california institute of technology DNA logic arrays.
Imminent: Superconducting quantum computers from Dwave systems. 16 qubits in Q1 2007 and probably 64 qubits by Q4 2007.