One of the things that struck me when I visited Northwestern and talked with Stoddart is how unbelievably developed the field of molecular computing is. Researchers are doing some compelling stuff with molecules in an effort to leave silicon and its intrinsic limits in the dust.
Molecular computing has a rich history dating back to the mid-90’s and before. In 1994, Leonard Adleman solved the “traveling salesman” problem using DNA. The problem asks whether, given a set of cities (“vertices”) and a set of paths connecting them (“edges”), a path exists which starts at a given vertex, passes through all the remaining vertices once, and ends at a second given vertex. Each piece of edge DNA complemented half of one vertex DNA and half of another, so that a “soup” of vertex and edge DNA easily provided every possible combination of edges and vertices. A series of amplification and purification steps then yielded the DNA containing the solution path. This was one of the first examples of encoding information directly in molecules, although the interpretation of the results was mostly human.
DNA computing research continues, but with the advent of topological chemistry and “molecular machines” (not to be confused with “nanocrap”), some researchers have turned their attention towards emulating electronic devices using molecules. Rotaxane-based systems, such as Stoddart’s 160-kilobit memory cell, rely on the use of bistable rotaxanes as a molecular switch holding a 1 or 0. You can imagine the dimensions of these molecular switches are smaller than even the teeniest of silicon transistors.
Stoddart’s cell, which corresponds to about 20 kB of RAM, has a density of memory elements that Moore’s Law projects for conventional memory in 2020! Of course, memory is useless without anything to process the information it holds, and as a result several groups are working on molecular logic gates these days. The basic idea of these is a change in the spectrum of a chromophore upon the addition of some “input” binding ion or molecule. De Silva’s XOR gate is a “push-pull” chromophore that binds calcium(II) and hydrogen ions such that if either one is present a spectal shift occurs, but if both are present their effects cancel and the result is no shift! Angew. Chem. Int. Ed. has a fascinating review on the subject (46 (22), 4026), and Tian et al. recently published on their development of a half-subtractor (a subtractor of bits) based on this same charge-transfer principle.