Let’s take a look at a new paper that drags in several topics that have come up on this site over the years – Birch reductions (particularly trying to get them done without the classic liquid ammonia/alkali metal conditions), continuous flow chemistry, electrochemistry, and more. The Birch has always been of interest for its ability to “dearomatize” rings – no, not to keep them from smelling, that’s rather in the sense of breaking up a conjugated delocalized aromatic ring (benzene and the like). This is not an easy process, since such systems are energetically favorable, but it allows you to get to some really interesting and useful intermediates. But doing it on scale is something that people would rather avoid, since you need an awful lot of liquid ammonia (at -33C or so) and plenty of sodium or lithium metal.
In the end, the Birch is all about transferring electrons into your starting material, and it works because the alkali metals have single outer-shell electrons that they’re ready to give up, and because ammonia can actually dissolve free electrons sitting right out there in solution (which is the source of the famous deep blue color of Birch conditions). But there are other ways to move electrons around, and one of the most direct is straight-out electrochemistry: bring your starting material into contact with a metal electrode and just pump ’em in there, not to get too technical about it. But once you’re past that conceptual stage, you do kind of have to get technical about electrochemistry, because there are a lot of choices in electrode material, current, voltage, solvents, electrolytes dissolved in those solvents, divided versus unitary cells, and on and on. Recent years have seen some deliberate attempts to make all this experimentally more accessible and reproducible, and it does seem to have led to an uptick in electrochemical methods.
The paper linked to in the first paragraph is part of a series from a group at Nottingham who are interested in both electrochemistry and flow methods. It’s a nice look at some of the engineering that comes into play when you start to scale these things up, and one of the biggest concerns (in flow chemistry or batch) is always mixing. That sounds like a rather low-tech concern, but on scale it can truly be a make-or-break issue. We early-stage bench chemists rarely have to worry about it, because on our tiny scales everything we do is probably mixed way more than it even needs to be. That’s easy to achieve in a 2 mL reaction volume, but while you can take the reaction vessel up to 2000 liters, you cannot scale up the little flask-spanning magnetic stirring bar in an equivalent manner – for one thing, if you tried it, you’d probably run into cavitation in the solvent.
Efficient mixing becomes all the more important when you’re having to expose your dissolved reactants to a continuous solid surface like the metal electrodes needed here. The apparatus described is inside an aluminum tube, which acts as the (sacrificial) outer electrode, with a steel rod running through it as the inner electrode. But it’s not just sitting there. The inner electrode is rotating rapidly, with the space between it and the walls adjusted carefully to produce “Taylor vortices” in the liquid. This is a complex subject in fluid dynamics, but it can be pictured pretty easily. If you get the geometry right, you end up with what’s more or less a stack of doughnuts inside your reactor. Each of these is a torus of rotating fluid, rolling along in the way that a thick circular rubber band or O-ring does as you roll it down the length of a rod. Each of the torii in the reactor are rotating independently of each other – in fact, each one is going the opposite direction of the ones next to it. But they are sort of brushing up against each other, and as the flow goes down the space between the tubes, things can get sort of handed off from one to the next. That gets complicated pretty quickly – depending on the size and spacing of the two annular tubes, the viscosity of the liquid, the rotation rate of the inner tube, and the flow rate involved, you can get all sorts of flow regimes: the vortices might stay pretty much in place, or move downstream themselves at some other rate, and the solutes might stay more or less in each individual rotating donut or be passed between them. You can get oscillating patterns in the size and distribution of these vortices (laid on top of the initial pattern), or they can take on a wavy appearance that can break into real turbulence, etc. That last link will tell you a lot more, if you’re interested, but that take-home is that these sorts of flows can do an extremely thorough job of mixing if you understand the system well enough to take advantage of them.
And that’s what the reactor described in this flow electrochemistry paper does. Running napthalene in at one end in THF as solvent, along with added dimethylurea and lithium bromide, spits out either the “mono-Birch” or “di-Birch” products (one ring dearomatized or both), depending on how you set the current. The authors demonstrate around 80g/day production with what is not a very large reactor. You do have to filter or centrifuge out some fine aluminum precipitate, since the outer tube is indeed corroded as the reaction goes on, and fortunately the design of the appartus seems to have kept it from clogging along the way (which is the absolute bane of every flow chemist’s existence). Folks are going to have to decide if that’s a deal-breaker, too.
All this took some careful attention to the fluid dynamics, as is the case with every attempt to use Taylor-vortex mixing. This paper did some computational modeling to start off with, and then checked that with real-world data from inline FTIR and Raman spectroscopy to monitor the reaction mixtures as they came out of the rotating tubes. For the right sorts of high-value products, this might work out quite well. But as with all process chemistry, there’s a committment step involved. Any new technology is going to have its own challenges and idiosyncracies, and you’re going to have to sink money and effort in at first to deal with those, in the hopes that everything can be tamed to the point that you can rev up the equipment and watch your desired product start coming out the other end. And that had better happen the same way every time you throw the switch!
