This week I noticed that J. Derek Woollins, head of Chemistry at the University of Saint Andrews and legendary selenium chemist, recently published a review of the research leading to the discovery of his eponymous reagent in Synlett. Although, for reasons of self preservation, I tend to lose interest in the chalcogens after the first row or two, I quite like personal accounts of research so I thought I’d give it a try. Unfortunately, being an organic chemist, I couldn't follow a lot of the inorganic chemistry but I did enjoy some of the stories of things going wrong:
“Paul Kelly also later extended this work by making M–Se–N complexes using tetraselenium tetranitride (Se4N4) as the starting material. This is a very sensitive material. Indeed the first time we prepared it (from a reaction in liquid ammonia), whilst washing the red solid with a solution of potassium cyanide, nitrogen was let into the flask fairly briskly. The resulting turbulence caused an explosion, which destroyed the flask embedding pieces of red selenium into the white shirt Paul was wearing; from a distance this looked like major bleeding. Bravery was not lacking in the group during that era, and Paul Kelly also carried out what remains as the only published reaction of (extraordinarily explosive) pentasulfur hexanitride (S5N6) to give a complex containing the [S2N3]– ligand.”
Ah, wacky fun! I’m quite sure that S5N6 falls into the category of things that shouldn’t really exist, and trying to force them to do so isn’t good for anyone’s nerves. Actually, the group engaged in quite a lot of this kind of lunatic chemistry. Another great example:
“Leaning on our previous work on reactions in liquid ammonia, we prepared sodium selenide (Na2Se) by the simple reduction of selenium with sodium in liquid ammonia. The resulting material is much more soluble than that prepared by the solid state route, though it is worth noting that this wonderfully finely divided solid is also very pyrophoric – on one occasion, around 75 grams caught fire in the port of our glove box with rather unpleasant consequences.”
I imagine that ‘rather unpleasant’ probably doesn’t cover it. I can think of few things that would evacuate a building faster than screaming the words ‘selenium fire!’. Fortunately, my own experience of selenium is a lot less exciting - I used phenylselenyl bromide a bit during my Masters to make enones, and aside from a marked decrease in affection from my girlfriend at the time, I suffered few ill effects. In fact, I suspect that beyond the old trick for forming enones (and the related Grieco olefination), many organic chemists would struggle to even think of uses for selenium in the lab. I am aware of one more reaction you can do with it, though, and I'll quickly explain why you shouldn't bother.
I'm sure most organic chemists reading this probably remember Baran’s excellent taxadiene synthesis from last year. Well, they say that there is nothing new under the sun, and one of the key steps - the intramolecular Diels-Alder to form rings A and B - was borrowed from Bob William’s 1995 route to the same target (albeit with a fair bit of optimisation).
However, although the groups shared a common tactic for formation of the A and B rings, their strategies for preparation of the above substrates were very different, as the disconnections below hopefully illustrate. Whereas Baran introduced the complete diene unit by addition of the corresponding organocopper reagent to vinylcyclohexenone early in his route, Williams constructed it in a more conventional stepwise manner later on. Unfortunately, this necessitated disconnecting across the tetrasubstituted double bond, which readers with a bit more synthetic experience will probably agree is usually something to try and avoid, as common off-the-shelf olefinations are notoriously poor at accessing this type of system.
Short of adding fairly hard organometallic reagents and following up with an elimination, there aren’t many general ways of forming these kinds of tetrasubstituted olefins, especially when there’s not much other functionality around. A few clever tricks like performing a McMurry reaction with acetone, or extruding CO2 from a suitable β-lactone have been reported, as has a truly terrifying selenium based olefination. As you may have guessed from the theme of this post, Williams decided (for reasons not given) to go straight to the latter. The reagent required for this step (which I think of as being a bit like a seleno-Peterson) is the dimethylselenoacetal of acetone. This is then treated with butyllithium to generate the corresponding selenium-stabilised carbanion, along with an equivalent of volatile selenide and strong feelings of enmity from your co-workers. Finally, addition of this carbanion to your carbonyl followed by elimination then gives your desired olefin (often in excellent yield), accompanied by more horrific selenium by-products and violent nausea.
A friend of mine once carried out one of these olefinations on a small scale in a separate lab on a different floor to where I work, and I still knew about it almost immediately, joining most of the department on an unplanned four hour tea break while the smell dissipated. Apparently the stacks on our fumehood exhausts aren’t high enough to disperse stuff that smells this bad, and it was all just getting blown out the top of the building and then wafting back in through the windows. Hell, doing the olefination step isn't even the worst part - you didn’t imagine than Aldrich sold that selenoacetal, did you? Nope, you've got make that yourself, although I hear it's easy to synthesise from acetone and selenomethanol (if you can ignore your co-workers throwing up on the other side of the lab, and you don't mind people crossing the street to avoid you as you walk home). My friend actually went the whole hog and even made her own selenomethanol, by gently refluxing selenium pieces in dimethyl sulfate, followed by reduction of the resulting dimethyldiselenide and careful distillation. In the end, the reaction didn’t work very well on her system, and she was forbidden by the section safety officer from ever running it again, but at least she always wins when the inevitable ‘nastiest chemistry you’ve ever done’ stories come out.
1. For another recent example of some daredevil chemistry, see this paper on the first preparation of azidoacetylene in Angewandte Chemie from a couple of weeks back.
2. Selenium, of course, takes its name from the Greek σελήνη, meaning moon, where the rest of us would prefer that all future research on it is conducted.
3. My own summary and views of Baran's route can be found here.
4. It's tempting to look at Williams' route (J. Org. Chem., 1995, 60, 7215) next to Baran's and dismiss it as a bit crap as it's twice the length, and racemic. However, it's important to remember that it was designed with late-stage incorporation of isotopic labels at key positions in mind, which no doubt made retrosynthesis a bit harder.
5. Seriously. For example, in Gerry Pattenden’s synthesis of isoamijiol:
I once heard Gerry describe this yield as ‘ridiculously good’, a term that (although absent from the paper) is probably fair. They couldn't even add isopropyllithium to that sucker.
6. This actually did happen.