Did Smalley deliver a killer blow to Drexlerian MNT?

The most high profile opponent of Drexlerian nanotechnology (MNT) is certainly Richard Smalley; he’s a brilliant chemist who commands a great deal of attention because of his Nobel prize, and his polemics are certainly entertainingly written. He has a handy way with a soundbite, too, and his phrases “fat fingers” and sticky fingers” have become a shorthand expression of the scientific case against MNT. On the other hand, as I discussed below in the context of the Betterhumans article, I don’t think that the now-famous exchange between Smalley and Drexler delivered the killer blow against MNT that sceptics were hoping for.

For my part, I am one of those sceptics; I’m convinced that the MNT project as laid out in Nanosystems will be very much more difficult than many of its supporters think, and that other approaches will be more fruitful. The argument for this is covered in my book Soft Machines. But, on the other hand, I’m not convinced that a central part of Smalley’s argument is actually correct. In fact, Smalleys line of reasoning if taken to its conclusion would imply not only that MNT was impossible, but that conventional chemistry is impossible too.

The key concept is the idea of an energy hypersurface embedded in a many-dimensional hyperspace, the dimensions corresponding to the degrees of freedom of the participating atoms in the reaction. Smalley argues that this space is so vast that it would be impossible for a robot arm or arms to guide the reaction along the correct path from reactants to products. This seems plausible enough on first sight until one pauses to ask, what in an ordinary chemical reaction guides the system through this complex space? The fact that ordinary chemistry works one can put a collection of reactants in a flask, apply some heat, and remove the key products (hopefully this will be your desired product in a respectable yield, with maybe some unwanted products of side-reactions as well) tells us that in many cases the topography of the hypersurface is actually rather simple. The initial state of the reaction corresponds to a deep free energy minimum, the product of each reaction corresponds to another, similarly deep minimum, and connecting these two wells is a valley; this leads over a saddle-point, like a mountain pass, that defines the transition state. A few side-valleys correspond to the side-reactions. Given this simple topography, the system doesnt need a guide to find its way through the landscape; it is strongly constrained to take the valley route over the mountain pass, with the probability of it taking an excursion to climb a nearby mountain being negligible. This insight is the fundamental justification of the basic theory of reaction kinetics that every undergraduate chemist learns. Elementary textbooks feature graphs with energy on one axis, and a reaction coordinate along the other; the graph shows a low energy starting point, a low energy finishing point, and an energy barrier in between. This plot encapsulates the implicit, and almost always correct, assumption that out of all the myriad of possible paths the system could take through the hyperspace of configuration space the only one that matters is the easy way, along the valley and over the pass.

So if in ordinary chemistry the system can navigate its own way through hyperspace, whats different in the world of Drexlerian mechanochemistry? Constraining the system by having the reaction take place on a surface and spatially localising one of the reactants will simplify the structure of the hyperspace by reducing the number of degrees of freedom. This makes life easier, not harder surfaces of any kind generally have a strong tendency to have a catalytic effect but nonetheless, the same basic considerations apply. Given a sensible starting point and a sensible desired product (i.e. one defined by a free energy minimum) chemistry teaches us that it is quite reasonable to hope for a topographically straightforward path through the energy landscape. As Drexler says, if the pathway isnt straightforward you need to choose different conditions or different targets. You dont need an impossible number of fingers to guide the system through configuration space for the same reason that you dont need fingers in conventional chemistry, the structure of configuration space itself guides the way the system searches it.

This is a technical and rather abstract argument. As always, the real test is experimental. There’s some powerful food for thought in the report on a Royal Society Discussion Meeting “‘Organizing atoms: manipulation of matter on the sub-10 nm scale'” which was published in the June 15 issue of Philosophical Transactions. Perhaps the most impressive example of a chemical reaction induced by physically moving individual reactants into place with an STM is the synthesis of biphenyl from two iodobenzene molecules (Hla et al, PRL 85 2777 (2001)). To use their concluding words “In conclusion, we have demonstrated that by employing the STM tip as an engineering tool on the atomic scale all
steps of a chemical reaction can be induced: Chemical reactants can be prepared, brought together mechanically, and finally welded together chemically. ” Two caveats need to be added: firstly, the work was done at very low temperature (20 K) presumably so the molecules didn’t run around too much as a result of Brownian motion. Secondly, the reaction wasn’t induced simply by putting fragments together into physical proximity; the chemical state of the reactants had to be manipulated by the injection and withdrawal of electrons from the STM tip.

Nonetheless, I rather suspect that this is exactly the sort of reaction that one would say wasn’t possible on the basis of Smalley’s argument.

(Links in this post probably need subscriptions).

3 thoughts on “Did Smalley deliver a killer blow to Drexlerian MNT?”

  1. Smalley delivered a killer blow to Smalley! His documented ignorance about two decades of anhydrous enzyme research–which falsifies a major point in his argument–should make anyone think twice about trusting him on this subject.

    Another way to look at the reaction-space argument: Picture a swamp, where pools of mud correspond to undesired reactions. To make the desired reaction happen, you have to cross the swamp. Smalley says Drexler’s robotic approach (controlling only a few degrees of freedom) amounts to trying to find a straight-line path across the swamp, and (Smalley claims) the pools overlap from a line-of-site point of view so that you have to twist and turn, using more DOF than a robot can supply. The “hydrogen-bonding genius of water” supposedly supplies the extra DOF to make enzyme chemistry happen.

    But we know Smalley is wrong about the function of water, so it may be that the extra DOF are unnecessary. In fact, according to Klibanov (the anhydrous enzyme expert), sometimes the water causes problems and enzymes work better without it. In addition, by doing reactions on stiff surfaces, Drexler has drained the swamp. There should be fewer side-reactions possible, because they now require not just a conformational change but a bonding change (rearrangement) with higher energy barrier. With fewer atoms in the picture, the dimensionality/complexity of the problem is reduced, and the robot looks more than adequate.

    By the way, Richard, what do you think of the recent report of single-bond bearings allowing free rotation in an engineered cage? http://pubs.acs.org/cen/news/8236/print/8236notw1.html

    It’s worth noting that the molecule doesn’t reconstruct; that the length of the chains can be engineered; that the rotor rotates rapidly at room temperature (presumably from Brownian motion) even with an asymmetrical rotor. It’s unclear from the news story how much of the rotor support comes from the cage/rotor interaction and how much from the Fe-P single bonds. But this certainly suggests that rotor/cage structures can be viable mechanical components.


  2. Chris, the work you cite is very nice, very much in the spirit of the sort of thing Fraser Stoddart has been doing with rotaxanes and catenanes. I think the challenges these kind of approaches need to overcome before they can do anything useful are firstly to demonstrate that their synthetic methodologies successfully scale to more complex architectures, and secondly to demonstrate a way of interfacing the molecules to the environment at the level of a single or a few molecules. (Currently most characterisation of them is done at an ensemble level in solution).

Comments are closed.