Digitizing present and past.
20 December 2024 - Wayback chemistry
Digitizing the chemical reaction spacing is progressing. The database has arrived in 1999 and a continuous set of organic reaction from USPTO patents from 1976 to 2024 is almost there. Digitizing the academic space pales in comparison with only piecemeal additions. This part of the project relies on publications that lend themselves to digitalisation of which there are few if you want to confine yourself to the present.
Will it be different if you rewind back to the past? A very useful resource is the journal Organic Syntheses, which is all about synthetic procedures and importantly, is also open-access. The focus is also on synthesis of compounds that are considered useful or synthesised using relevant procedures. The articles in general respect systematic names making them suitable for semi-automated digitization, the only disadvantage is that the actual procedures are too detailed and too long for even chatgpt to make sense of. In any event, the journal stretches back to 1926 and procedures for the synthesis of allyl alcohol, benzylic acid and benzoin from that year have already been incorporated.
And travel even further back in time? It is questionable how material from the nineteenth century and older is helping a dataset constructed for present day machine learning but it can still be entertaining. In a 2011 episode of this blog the news was welcomed that the Royal Society had open-accessed the Philosophical Transactions of the Royal Society (Link). The journal goes back to 1665 which may be over the top but for the purpose of exploratory research an article was selected at rondom from 1866 by Edward Frankland and Baldwin Francis Duppa (Link)
The article is about ether synthesis and has one easily recognizable synthetic procedure: add 6000 grams of sodium acetate, 3600 grams of ethanol and 9000 of sulfuric acid and get 6000 gram of "acetic ether". The acetic ether of course is ethyl acetate and the article has less to do with our modern concept of ethers and more with esters. The exactly 6000 grams of product yield seems suspiciously accurate, other data from the article (elemental mass, vapour pressure) are reported with all the sig figs. Another peculiarity is the recommendation to leave the mixture of ethanol and sulfuric acid, called "sulphovinic acid" (contraction of sulfur and wine) for 24 hours before adding it to the sodium acetate. Does mixing ethanol and sulfuric acid do anything except generating heat?
The article continues with much more chemistry but chemistry that requires some deciphering. The "acetic ether" is added to sodium metal and with ethyl iodide the endproduct is presented as "ethylic diethacetone carbonate". The elemental analysis (C10H18O3) is helpful and works out formation of first ethyl acetoacetate which is not isolated and then the di-ethylated adduct. The liquid has a very fragrant odour and an "aromatic taste" (please, don't) and when boiled in an ethanol solution of potassium hydroxide it forms sodium carbonate and a new compound presented as "ethylated acetone" but more recognizable as 3-ethyl-pentan-2-one. Incidentally, the synthesis of ethyl acetoacetate is supposed to be the first ever Claisen condensation, invented 21 years later in 1887?
Returning to the main topic of the blog, digitizing chemical reaction space of the past if the present does not have enough source material to work with. Include the Philosophical Transactions of the Royal Society for digitizing? No! Digitizing Organic Syntheses? Yes!