Lazy chemistry

Lazy chemistry
14 September 2011 - Orgo

Here is an ongoing trend in organic chemistry: how to make as much compounds as possible or try out as much reactions as possible with the least possible effort, preferably automated. Makes sense of course. Organic synthesis is time consuming and the more compounds available for lets say drug screening the better.

Take Robbins and Hartwig. In a recent publication they admit that many new reactions are stumbled upon rather than cleverly designed and in an ambitious screening program they demonstrate how to maximize the number of reactions a human can reasonably do without investing any time in reasoning how reaction can come about (DOI).

First they mixed together 17 organic compounds that each contain one functional group, among them 2-cyanonaphthalene, diphenylacetylene and dodecane. This mixture was distributed over 384 positions of a microtiter plate. Then to each of 16 horizontal rows was added an excess of a metal compound (among them CuCl, FeCl3 and MoCl5, one blank) and to each of 24 vertical rows was added a ligand, for example triphenylphosphine, cyclooctadiene and diaminocyclohexane (and a blank). After a given reaction time the content of each well was analyzed by GC-MS with each potential product identified by the molecular ion peak. The number of potential reactions uncovered this way (also allowing for homocoupling) is 16 x 24 x 17 x 17 / 2 = 55,488.

The Robbins / Hartwig paper does not actually reveal the number of reactions discovered but highlights a few. One of them is the coupling of 1-dodecyne with 4-butylaniline catalyzed by copper chloride and not even demanding a ligand at all (sobering result). The added bonus of this high-throughput method is an estimate of functional group tolerance with all these other 15 compounds are just hanging around.

More lazy chemistry is reported by Osowska and Miljanic (DOI). Their target is the synthesis of new imines and they do this with a seemingly shotgun approach. In it 5 aldehydes and 5 anilines were dissolved in toluene in a single vessel and refluxed for 72 hours. In ordinary circumstances the reaction product would be a hopeless mix of 25 different imines impossible to separate but there is a twist. The reaction product, described here as an orange sticky solid was subjected to distillation in vacuo and after 4 days (mind you, this is proof of concept chemistry, not industrial-grade chemistry) a first fraction could be scraped off from the condenser which turned out to be surprisingly just one of 25 imines in 77% yield with 96% purity. Five days later the second fraction contained a second pure imine and many days later with increasing the temperature to 240°C the final 5th imine rested on the bottom of the flask resisting any further distillation.

The trick is that the most stable and lowest boiling imine will distill first. All the other imine forming reactions are reversible and in the spirit of Le Chatelier's principle the most stable imine will continue to be generated at the expense of the other imines. this is called self-sorting (see: dynamic covalent chemistry). When the two imine reactants are depleted, the next stable imine gets a go. In this system equilibration is based on thermodynamic reaction control and removal from the system based on kinetic control.