Expanding the scope of available reactions

The benefits chemists most often discuss in continuous flow methods involve: handling reactive or dangerous reagents, how more efficient the mass to heat transfer is, the ability to use in-line column to scavenge or quench a reaction, but something that has opened my eyes is the reaction space that is opened and how under-utilized it has been for a great number of years outside of some need to add a step in a total synthesis or specific methodology development……that’s right, I am talking about photochemical processes, reactive gas additions, ozonolysis, electrochemical methods and in terms of mixing — how to handle a higher percentage of solids in a flow reactor. An all important topic and it amazes me that in 20+ years in pharma, that I ran very few of these outside of a ozonolysis and some reactive gas addition chemistry (NH3/CO2/CO/H2) so it gives me an opportunity to pull off some of my dusty books off the shelf and have a look at things that are fresh and current again. For a start I thought I would highlight a couple of applications: a photochemical [2+2] cycloaddition (I see the handwaving now) and a reactive aromatic substitution with NH3 under high temperature and pressure as a starting point and then add some uncommon reaction methodologies and see if it gains some traction in the literature (enabling).

For the aromatic substitution with NH3, an application note provided by ThalesNano with a substitution of 3,5-difluorobenzonitrile with ammonia (as well as a series of amines, single and double subsitution) illustrates the ability to use NH3 and operate at very high temperatures and pressures under a continuous flow stream (X-cube). Just for perspective, without the aid of pressure or transition-metal catalyst, substitutions of F with an amine or ammonia on an aromatic ring with EWGs is a pretty sluggish reaction to say the least, with reactions in the neighborhood of 2-4 days in many reports. Microwave enhancements certainly offered a pretty good alternative technology for making this acceptable for a medicinal chemist to generate advanced substrates, but usually avoided the use of ammonia. Although the target compound can be made in different ways, the first proof-of-concept run entailed the flowing pre-absorbed NH3 in NMP and the 3,5-difluorobenzonitrile at 0.5ml/min into a 8 ml loop (16 min residence time at a temperature and pressure of 275C and 200 bar, respectively to provide the desired mono-substituted compound in >99% yield. The high temperature and pressure capability made this a tremendous opportunity to implement reactions that were otherwise unavailable in traditional set-ups.

Screen Shot 2014-10-31 at 1.38.35 PM

Following the success of the the initial run, the group extended to method to include additional amines. One noteworthy observation made had to do with solubility and in order to provide a way to continuously flow all substrate products and amines, a 6% MeOH solution in NMP was found to be the optimal solvent conditions used.

Screen Shot 2014-10-31 at 1.50.27 PM

And as you can imagine, why not add a second substitution to the additional F and round out a series of diamino-substituted benzonitriles, but one can imagine using other substrates to generate mixed substitution patterns as well. Several examples and a fully detailed account of the conditions are available in the note.

Screen Shot 2014-10-31 at 1.55.24 PM

Turning our attention to photochemistry – and I can count the number of these I have performed on both hands so bare with me, lol! Well that’s not entirely true, I dabbled a bit on the microwave side and the results were variable — in the microwave you generally apply and electrodeless lamp that is capable of discharge, and the power has to be just right. It does work but it takes some time….so to see the ease at which a lamp can be placed in the flow of reactants wets my appetite.

Screen Shot 2014-10-31 at 6.14.31 PMScreen Shot 2014-10-31 at 6.14.45 PM

In an application note #36 by Vapourtec, they set the stage by immediately talking about the chemical space that is now available by applying a UV-lamp (take a look at the set-up on their website) in line with the flow of reactants — no doubt this will get used extensively. They chose in this particular example a [2+2] photochemical cycloaddition of maleimide and 1-hexyne to show how a reaction can be set up effectively — because the flow of reactants and the photons/min are both important factors for a successful application. In fact they go on to point out that the photo cell UV-150 can be thought of a reactant that is simply not in-line with the reagents. Man this brings me back to a book every organic chemist should own (Organic Chemistry: The Name Game: the fun part is in the nearly all carbon nature of the chemistry discussed – and photochemistry comes to mind). So back to the reaction: Since the UV-lamp is emitting photons in a specific range, the amount needs to match the transformation and the flow or residence time in front of the lamp.

The flow reactor schematic in the E-Series is shown below — the coil tubing is placed around the photoreactor. And I should point out that photochemical reactions are not thermally driven so in their case they cooled the reactants coming through to 30C (but studied and optimized for higher temperature to make sure). In addition, different filters were used to keep the range of photons at the right frequency for the reaction.

Screen Shot 2014-10-31 at 5.59.12 PM


With an initial reactant concentration of 0.1 M , the flow rate was studied at 1-8ml/min with a constant UV at a constant 120W. The result indicated that 4 ml/min provided >95 yield of the desired. By increasing concentration to 0.4 M and slowing the flow rate to 1 ml/min they could achieve the same conversion, and thus the same output/day. The scheme is shown below but I encourage you to take a look at the application note; it is extensively studied, with a number of details that will help ease you into this new area of continuous flow chemistry and unexplored territory. Happy Reading!

Screen Shot 2014-10-31 at 6.07.56 PM