|
Positional Synthesis vs. ChemistryIn November of 1993, there was a meeting in Santa Monica. Bill Goddard, who was physically present, discussed his work with self-assembling molecules. This was a shorter version of the presentation he had just given at the Foresight Conference. Ralph Merkle, who was in San Francisco, addressed the group via a closed circuit video hookup. We could see and hear him, he could only hear us. He said it surprised him that it was possible to make molecules self-assemble -- "It's as if you took the parts of a radio, put them in a bag, shook the bag, and got a radio." In the following discussion I said this shouldn't be surprising, because that's how chemistry works. There is nobody saying "you go here," etc. It's like the difference between central planning and a free market... At this point there was a collective gasp from the audience, and Ralph did a double-take. In this section I want to elaborate on what I meant. When you have a trillion things that are supposed to work in step with each other, sending explicit instructions to each one is basically not a good idea. In economics, you can either have somebody in Moscow directing everybody -- "You go here and do this, you go there and do that" -- or you can let people explore their local environment and make their own decisions. At the atomic level, you can either have nanotechnology, in which atomic positioners pick atoms up and put them in a certain place ("You go here... ") or you can have chemistry, in which the atoms are guided into place by their local environment. In chemistry, atoms move in three dimensions. An atom is a multi-faceted thing which lives in an extremely rich environment and interacts with other atoms in an intricate dance. Atoms are constantly vibrating and bouncing around. Sometimes they find a place where they fit (i.e. they form a compound or crystal) and stay there until the continual interatomic jostling (heat) knocks them loose again. In other words, molecules self-assemble using local information. This process of self-assembly is analogous to a self-organizing economy. Let's take a very simple example. Suppose you are running a country, and you want to make sure that the economy produces plenty of salt. Your desired objective is to make sure salt is available at a reasonable price to everyone who wants it. To make that happen, which is better -- to let producers make salt in a free economy, or to have a bureaucracy directing everyone's activities? The price mechanism gives producers, middlemen, and consumers all the information they need. You can ignore this natural economic information and try to use central planning instead, but... people will have to stand in line for hours to get their monthly ration of salt. Now, suppose you want to make a grain of salt. Which is better -- to start with a solution of sodium and chloride ions, and then evaporate the water, or to have 10^12 atomic positioners putting atoms together? How do you coordinate the activity of 10^12 atomic positioners? If you just let the water evaporate, you don't need to coordinate anything. The ions form salt crystals all by themselves. They have all the information they need. You can ignore this natural chemical information and try to use digital information instead (i.e. a digital model of where everything is and what needs to be done), but this will make your task a lot harder. If you just get out of the way and let it happen, all the sodium ions and all the chloride ions will automatically position themselves through a natural diffusion process so they can react with each other. Why would it be more efficient to pick them up one by one and put them into place? Can anyone explain (or calculate) why making salt with atomic positioners would be faster and/or cheaper than letting crystals form in the natural way? A few years ago an engineer named Pravin Mistry accidentally discovered a technique he could use to apply a coating of diamond to a substrate. If you put an object in an atmosphere of carbon dioxide, and shine lasers on it, "the intense heat at the spot where the lasers converge breaks up some of the gas into an electrically charged plasma. The lasers also vaporize a very thin layer of the object being coated..." Atoms from the hot gas are deposited on the coated object, forming diamond. Can anyone explain (or calculate) why making diamond with atomic positioners would be faster and/or cheaper than the Mistry process? The same question can be posed about smart materials, or any materials. If you can make some material or some object with chemistry (in which I include bulk processes such as the Mistry technique), why would it be faster and/or cheaper to make it with positional synthesis? Long before the Soviet economy collapsed -- before the Soviet Union even existed -- economists had been analyzing the idea of central planning and arguing that it would never work. According to Davidson and Rees-Mogg (Blood in the Streets, p. 168), this idea was originated by a Dutch economist, N. G. Pierson, in 1902. By the 1970s economists had arrived at a mature theory showing that when you give up the price mechanism, you give up most of the information you need to run the economy. By the 1990s this idea had penetrated the public consciousness almost everywhere. Even in most third world countries, governments have given up on central planning. Nevertheless, there is still a deeply held belief (usually an unconscious belief) to the effect that planning must be more efficient than self-organization. Many people still think, or feel, that if we could just control everything, the world wouldn't be such a mess! This belief or feeling is wrong, of course. Both in economics, and on the atomic level, it's usually better to let events organize themselves using local information, rather than having a central authority directing events according to its model of the situation. There are exceptions to this. In wartime, when nations absolutely have to maximize production in the short term, they forget about free markets and say "You go here and do this!" It works for a few years. It also works in corporations. Management does tell employees where to go and what to do, according to its model of the situation. Central planning will probably turn out to have uses on the atomic level as well. Some structures (nanocomputers are the obvious example) may be beyond the capability of chemistry. If they are going to be made at all, they will probably have to be "assembled" with atomic positioners. (Not necessarily, however. Nanocomputers may not be beyond the capability of chemistry assisted by biotechnology.) I'm not saying that positional synthesis is impossible just because it involves central planning. What I'm saying is that central planning won't be used routinely. It's not a question of what can or can't be done with atoms. It's a question of what it would cost to assemble structures atom by atom, compared with the cost of making equivalent structures with chemistry. Positional synthesis will be an extension of chemistry, not a replacement for chemistry. Anything that can be made with bulk processes will be made with bulk processes, and anything that can be made with biotechnology will be made with biotechnology. Positional synthesis will be most expensive of the three possibilities, and therefore it will be the method of last resort. Atomic positioners won't be used to make common things like bricks, chairs, and vacuum cleaners. They will only be used to make things that could not be made in any other way. |