That outpost of empire Australia Produces some curious mamalia: The kangaroo rat, The blood sucking bat, And Arthur J. Birch inter alia
-Sir John Cornforth 
The Birch reduction is a useful, historic and powerful transformation, and has numerous variations and applications that expand its scope far beyond the simplest case taught to undergraduate chemists around the world. Liquid ammonia might seem like an exotic, dangerous (or even scary) thing to work with but, like all reagents, it can be used perfectly safely as long as the correct procedures are followed. This isn’t a step-by-step guide for running a Birch reduction, and is not a replacement for proper training in the use of liquid ammonia, but will hopefully provide a glimpse into the practical aspects of this chemistry for those with only theoretical knowledge of it.
Tip 1: Do some reading and find a good procedure
Go and read some Organic Syntheses procedures; they’re a goldmine of trustworthy information. The Org. Synth. procedure for reduction of Anisole was penned by Birch himself, and is the best example of a ‘modern’ style Birch reaction procedure I know of. Myers’ notes on the reaction (pdf) give a nice review of the theory and applications. It isn’t easy to predict which metal (Ca, Li, Na, K…) is best for your substrate, so have a look in the literature to find conditions used on something similar looking.
It’s also a good idea to talk to people, and find someone who’s done one to help you the first time. I watched a documentary a couple of years ago about a guy who wanted to take up the notoriously dangerous sport of BASE jumping; he simply borrowed a parachute from a skydiving friend of his, went out alone to a cliff with no idea what he was doing and jumped off it. He survived, just. Don’t be that guy .
Tip 2: Choose the right setup
Below is the setup I recently used for the small-ish scale (500 mg) Birch alkylation of an aryl ester with lithium as the metal. Obviously, your mileage may vary. The Birch alkylation is very similar to the Birch reduction; the main difference is in what is used to quench the anions generated in the reaction. In a Birch reduction, the reaction is normally done with a proton source (usually t-butanol or ethanol) present in excess, so that the carbanions formed during the reaction get protonated straight away. Conversely, in a Birch alkylation, only one equivalent of proton source is present, and this only mono-protonates the dianions generated in the reaction. This means that, in the case of aryl esters, you’re left with an enolate, which can be alkylated with alkyl halides etc. In this case I used iodomethane, but most alkyl bromides and iodides work fine. Surprisingly, liquid ammonia is a great solvent for alkylations!
As you might imagine, Birch alkylations are a bit harder to do as they’re much more sensitive to water and impurities in the ammonia, and they require considerably more care than the classic Birch reduction, which is actually a pretty robust and tolerant reaction. In the photo below I have my glassware assembled and cooling under nitrogen after a night in the oven. On the left is my drying flask where I’ll condense my ammonia and dry it with some sodium before distilling it over into my reaction flask on the right. More on this below (See Tip 4).
The ammonia inlet is just a short pipette stuck through a quickfit thermometer adapter. You can also use a needle and septum for this, as long as the needle has a sufficiently wide bore. DO NOT use a regular narrow bore needle here, unless you’re very confident in your ability to regulate ammonia pressure, and even then it will take ages. Ordinary condenser tubing is used to connect this flask to the reaction flask. I’ve use a quickfit tube adapter on the left and another pipette/thermometer adapter combo on the right.
Note: If you’ve never used it, ammonia comes in pressurised cylinders of various sizes; all the ones I’ve used have been large, heavy and trolley mounted. Sometimes you get the option of dispensing liquid or gaseous ammonia depending on how you orientate the cylinder. On our trolley mounted one you can ‘select’ which you get by rotating it. It doesn’t really matter whether you’re dispensing liquid or gas, although liquid is considerably faster and is better if you need more than 100 mL or so. Remember that liquid ammonia is a cryogen and often makes the tube connecting the cylinder to the reaction cold and brittle.
Personally, I do not use a trap to stop escaping ammonia. If I’m using my nitrogen line (as I am here), any ammonia that gets past the condenser goes through my (mineral oil) bubbler into the back of my fumehood. For large Birch reductions or the preparation sodamide I often just have a tube going to the back of the fumehood. In any kind of working fumehood this is fine.
Ammonia cylinders vary a lot. I’ve use a couple of different ones, and they all have really stiff taps that are difficult to open just a little bit. The first ammonia to go into your flask will just vaporise, causing a pressure spike that can make things pop out of place, even if they’re Keck clipped in place. For this reason it is absolutely vital that you clamp the end of the tube from the ammonia cylinder inside your fumehood. Someone in a lab I used to work in didn’t do this and opened the tap too far, causing the tube to fly out of the flask and spray ammonia all over the floor. The chemist was fine, but we evacuated the lab for a couple of hours while the smell dissipated.
Personally, I always have a three neck flask for my drying flask and remove the septum (see above photo) before opening the ammonia tap. Then there’s no risk of pressure spiking and popping things out of place (like an expensive dry ice condenser). Once you get the flow how you want it then you can loosely put this septum back. If I didn’t have a drying flask I’d probably take my condenser out the first time I opened the tap. You get the idea.
One really important tip for reproducible yields when using dissolving metals in liquid ammonia is to make sure that your ammonia is dry and iron free. Even trace iron from the inside of your ammonia cylinder can have a dramatic effect on the rate at which various metals dissolve, dramatically affecting the reaction. Ever see a procedure using freshly prepared sodamide (NaNH2)? Chances are, if it’s a good one, it’ll use a pinch of an iron salt to speed up formation of the base. However, if your aim is to have a nice blue solution of dissolved electrons, this is bad. This is ESPECIALLY IMPORTANT FOR SODIUM, which is much more affect by the presence of iron than lithium .
Using the reaction setup shown above I collected 50-60 mL or so of liquid ammonia in my first flask and dried it with a few small pieces of sodium:
The fact that the ammonia goes blue straight away shows that it’s already pretty dry, which is good, but I’ll still stir it for 5-10 mins before use.
This means that it takes a long time to evaporate . With a flask full of nice, dry ammonia it’s now time to distil the required amount into our reaction flask. The receiving/reaction flask is now cooled in a dry ice-acetone bath and the dry ice condenser is also filled with the same mixture. The cylinder is disconnected, the drying flask is removed from its cooling bath and… nothing happens! 
For this reaction I simply left the setup as it is in the photo and timed how long it took to distil the 40 mL of ammonia I needed across into my reaction. It was a little over two and a half hours. For this reason, particularly on large scale, I usually gently warm the drying flask with a heatgun at this point. This might seem silly, but it does expedite things somewhat.
Tip 6: Allow enough time
Once you’ve got substrate, ammonia, co-solvent and proton source in the reaction flask, things don’t take too long. These reactions are a little tricky to TLC (although it can be done), but fortunately, colour is a really reliable guide here. Most people know that metals dissolving in ammonia give an incredible dark blue solution (like the colour of my drying flask in the photo above). This is due to the presence of solvated electrons as the metal dissolves. While there’s still starting material left, all the electrons very rapid go into reducing that, so as you add the first pieces of metal, nothing really happens. Only when all the starting arene has been consumed will there be excess electrons to give the blue colour, and that means that the reaction is done.
By way of illustration, here’s a video of my Birch alkylation in action. I’ve got everything in the flask and I slowly add the lithium piece-wise. At first, not much happens, but the last piece causes the solution to change colour, indicating all my starting material has gone . I stir the reaction mixture for another 30 seconds or so and then add my electrophile (MeI) as a THF solution. The video is completely unedited (yes, really), and you can see that it’s all over in a few minutes, even though I get my last piece of lithium stuck to the side of the flask!
Of course setting up, drying and distilling, and warming up to room temperature afterwards can take hours, but at least you can do other stuff at the same time.
Tip 7: Finishing up
So you’ve successfully run the reaction, and you’ve just quenched it. Now what? Providing that your substrate is not air sensitive, and won’t mind a little longer in ammonia, the easiest thing to do is just leave the reaction in a cork ring in the back of the fumehood overnight for the ammonia to evaporate (See Tip 5). You don’t want to work it up with any ammonia in, because it will stink. If you’re not so confident in the stability of your product, you can remove ammonia in an hour or two just by blowing nitrogen through the reaction mixture. For the above reaction, I used another pipette in a thermometer adapter to pass nitrogen through my reaction mixture and had a piece of tube running from the other neck in my two-necked reaction flask (replacing the condenser) to carry the ammonia/nitrogen gas to the back of the fumehood.
If anyone has any questions or tips of their own, please get in touch in the comments!
1. I can’t find a reference for this limerick, so I hope I’ve remembered it right. I believe it was attributed to Birch’s countryman (and 1975 Chemistry Nobel Laureate) Sir John Cornforth.
3. Sodium is way more susceptible to iron impurities than lithium. By about 25 times. From Fieser and Fieser, page 55:
Regarding the superiority of sodium, I’ve usually found lithium to be at least as good, and easier to work with. My experience has generally been that lithium is much better for naphthalenes, which are mostly what I’m into at the moment.
4. A (possibly apocryphal) story I heard from an emeritus professor who knew Birch when he was at Oxford relates how he used to get his liquid ammonia from a nearby chemical company for free as part payment for some consulting he did. All he had to do was provide his own vessel to take it away in. Fortunately, as ammonia evaporates quite slowly, he used to just get a couple of large glass carboys (the kind you use to make wine) and fill them up with liquid ammonia. He’d then wrap them in cotton wool, balance them carefully in the back seat of his car and drive quickly across town with the windows wound down and the ammonia happily sloshing around behind him. Rather him than me.
5. Tip 5.5: Leave the end of tube from the ammonia cylinder in the fumehood for a good hour or so after you disconnect. The smell of ammonia is strong, lingering and unpleasant!
6. The metal is usually added a piece or two at a time, and will dissolve faster if the pieces are pressed flat using a spatula to maximise their surface area. It’s especially important not to use a large excess if you have other reducible things present in the molecule (e.g. esters, other olefins etc).