Just read an article (Molecules 2014) on flow through hydrogenations through the eyes of a medicinal chemist….sort of. A combination out of Baxendale and Ley is a contribution in the area of heterocyclic construction using an appropriately placed hydrogenation of an aromatic nitro group, strategically located to take place in a subsequent reaction to form advanced riboflavins, quinoxalinones and benzodiazepines…..each important in their place as strong pharmacophores.
Two things that stick out to me as important: how do they arrive at a final working method and what were the issues….when you read an article by these authors it will inevitably have this information present and that is the type of discussion that will help push the area of flow chemistry forward. For instance, in the first scheme below, the group needs a diamine functionality and over-reduction of an aromatic halogen took place with Pd/C but not with PtO2 (ha, how many times have you read a BIOMCL — the things that don’t work are not discussed)— and this is likely made more challenging by the fact that they had to heat the reaction to 45C for the reaction to work well and use MeOH to keep the materials in solution for the duration of the reaction. The nice thing about the reaction is that it was easily be performed and optimized for catalyst using the H-Cube by ThalesNano. Another piece of anecdotal information — the diamine is typically not stable for any appreciable length of time — but can be used in the subsequent step. The scheme below indicates the two possibilities for the reaction — and note that the condensation here was done in batch fashion for 10 hrs at room (must decompose the diamine with heat because 10 hours is a long time).
Although they applied the same methodology to a quinoxalinone series, I am going turn our attention to the work done on benzodiazepines. Work on the scaffold is traditionally considered older — with new developments as a rarity — I know my early research included an aza-version of this and it felt like a total synthesis. For the storyline, some conditions needed some attention in order to move to a complete flow system. For example to build the amino-nitro diarene, they chose a microwave mediated SNAR reaction of a fluoro-nitro arene and the requisite aniline. Fortunately, I have discussed this type of reaction in a past post so there is some development utilizing this approach — in this case, however, the aniline is deprotonated with LHDMS and irradiated in the presence of the fluoroarene to produce the diaryl scaffold in high yields in a short amount of time. The product was redissolved and hydrogenated and cyclo-dehydrated to provide the benzodiazepine in high yield. To adequately handle the dehydration, an in-line MgSO4 filled glass Omnifit column was placed subsequent to the flow hydrogenation (again – H-Cube with in-line MgSO4).
The great thing about the microwave method is that it provided a good reaction starting point for a medicinal chemist planning a library for some initial screening hits. The bad news is that scale-up in this fashion would require a different microwave or a continuous flow method if additional testing or to get the compound through an entire cascade an on its’ way to animal studies….this group recognizes this as a key criteria in developing the technology and therefore worked out a method for the first step in a continuous format. Prior to jumping into the flow conditions, the 2-step process provided a nice route into the desired compounds with microwave (1) and flow (2) method.
In moving over to the flow conditions, the base and mixing were critical for the execution. Although the scheme helps you follow the format, three separate lines were used to form a good process with n-BuLi in channel A, the amino-benzophenone in channel B and the fluoroarene waiting in channel C, with the appropriate mixing chambers and T or Y-lines adjusted to provide the mixing and timestamps for delivery of reactants/reagents. The initial solutions for deprotonation were cooled to 0C, mixed for a quick reaction and flowed into a flow stream of the electrophile. Once this last mixing is started, the flow went into a heated coil loop (52 ml) at 115C for the cyclization to proceed. Once the product was formed, the group added a process to quench and work-up the reaction so that the desired solution of organic product (plus the addition of MeOH if needed) would flow into the H-Cube midi under similar conditions indicated above (5 bar, 45C, 2.2 ml/min, coil loop, PtO2) provide up to 120 mmol or 38.1 g. The in-line work-up process is certainly worth a detailed read — this is an area of discussion that chemists will eventually develop innovative ways of handling reactions that include lithiations mid-stream (and other transformations)…although we would like to make everything plug-and-play, we need to make sure to understand what’s in the line to have a successful flow method in the lab. Happy Reading!