Skip to main content

Chemical News

Ah, Just Pour It Into Salt Water

Today’s post is one for my fellow organic chemists to wonder over. This new paper from a group at the University of Bari describes a palladium-catalyzed coupling reaction of alkyllithiums and aryl halides. And that in itself is not that remarkable – it’s not easy to get that combination to go, not least because you might end up with an alkyl halide and an aryllithium instead, but (as the paper itself takes care to note, the Feringa group has published a series of reactions in this category.

What makes this one unusual is the way the reaction is run: you take your aryl halide and palladium bis(tributylphosphine) catalyst and stir them in water with sodium chloride added. Then you run the speed up to a vigorous vortex, in the open air, and quickly add a solution of the alkyllithium in a nonpolar solvent (such as cyclohexane). In about 20 seconds you’re done, coupling in up to 98% yield. Yes, you’re adding an alkyllithium to water, and what’s more, you want to be sure to do it in the open air (deoxygenating the system lowers the yields). You need the chloride salt additive, too.

Now that’s a surprise. Most organic chemists picture the quenching of a reactive reagent like an organolithium as pretty much a diffusion-controlled process – the first water atom it sees gets reacted with. But apparently you can  get this coupling to go, in water, but get it to go in a Pd-catalyzed manner, which means that some of those organolithiums are having to wait their turn to react. The main way I can picture this happening is inside small droplets of the organic solvent. The relative concentrations of the reactants must be pretty high inside those and there’s a high surface area to aid in diffusion – you certainly can’t rule out some effect at the solvent interfaces as well, and if the droplets are small enough, they’re basically all interface. Meanwhile, the oxygen present is helping in some sort of oxidation process to keep an active catalyst, and the chloride is having an effect on the active catalyst structure as well (a phenomenon that’s been noted in more conventional reactions).

But it’s not a reaction that I would have thought to try! If you switch to methanol, all you get is dehalogenation of the aryl compound. If you try other Pd catalysts, the whole conversion breaks down. Secondary organolithiums work, as do heteroaryl halides. So if you’ve got a quick substitution to do and you have the lithium reagent on hand, give it a try – and come back here to the comments to tell everyone how it worked (!)

49 comments on “Ah, Just Pour It Into Salt Water”

  1. rhodium says:

    OK, I now officially know nothing about running reactions.

    1. Derek Lowe says:

      Yeah, that’s just the sort of feeling a publication like this brings on, isn’t it?

  2. Skippy says:

    This reminds me of one of my favorite “on water” reactions, the nucleophilic addition of alkyl lithiums and Grignards to imines and nitriles using water as the solvent. You basically just drip your alkyl lithium into a vortexing mix of water and imine and the reaction is done in seconds. They also observed that the reaction only worked in water, methanol and D2O both gave much lower yields.

    1. CuriousScientist says:

      The same group observing the same effect, they seems to be expert on water vortex unusual organic reactions. Surprising!

    2. M says:

      “Just tip the t-BuLi into an aqueous solution of your imine. And make sure it’s in an open flask”.
      “Erm… sure thing prof.”

  3. Jb says:

    Side note: ‘largest cartel in the history of the US’ – as reported by WaPo – might be involved in an elaborate price fixing scheme for nearly 300 generic drugs. 47 states are plaintiffs for which prosecutors say the generic industry is riddled with artificially raised prices. No wonder insulin prices are through the roof.

    1. Eugene says:

      And as with the Wall Street Crash of 2008-2009 there will be some token jail time for some fall guys and a series of fines. As soon as things settle down it’s back to business as usual.

      1. x says:

        Laws are only for the little people. Especially these days.

  4. tlp says:

    I wonder how that was discovered. Was it a bored grad student trying to quench BuLi in brine?
    Now my SN side reactions with t-BuLi don’t look that surprising.

  5. Nick K says:

    Astonishing finding, but I’m not sure I’d want to do this reaction on 10g, let alone a kilo scale……

  6. drsnowboard says:

    does it work in the southern hemisphere?

    1. Chris says:

      Only if you vortex in the counter-clockwise direction.

    2. El Tordo says:

      Nothing works down there… believe me!

  7. Heliastes says:

    “give it a try”… with caution in bold, underline, and italics :p

    1. oldnuke says:

      And don’t tell the safety people what you are about to do! 🙂

  8. Thatchemistguy says:

    They do have a safety warning in the supporting information stating that caution should be used on larger scales :).

  9. gippgig says:

    This sounds like it belongs in the “Things I Won’t Work With” cat egory.

    1. Scott says:

      I’m not even a chemist and it sounds like that to me!

      1. dave w says:

        Yeah – organic compounds of reactive metals are used in rocketry for “starting fuel”… not usually lithium, but things like alkyls of zinc, boron, or aluminum: apparently they react instantly with oxygen at any temperature, including liquid oxygen… they’re a convenient solution to the problem of starting e.g. a large oxygen-kerosene liquid rocket motor. You have two fluids spraying in at high flow rates, and one of them is very cold and the other is oily. You must set fire to them… VERY fast. It’s only dangerous if there’s a delay and you accumulate excess propellant in the chamber before ignition. So you include a small tank of the metal-organic stuff that gets injected at the beginning of the fuel feed, before the kerosene.)

        1. dave w says:

          I thought the usual laboratory protocols for using such things in reactions included closed systems and inert gas blankets… the salt-water reaction sounds a little hairy, but perhaps the same sort of situation: it’s only dangerous if you noodle around and don’t get everything mixed really fast…?

  10. K says:

    I’m not buying this until it has been reproduced in another lab.

    1. bob says:

      As a tech during my undergrad years I was running multiple preparations of boronic acids from electron rich phenyl bromides using nBuLi. Sometimes the principle products of the reaction were debrominated phenyl and nBu alkylated phenyl. My lab ran endless number of Suzukis at the time, and who knows, I might had some Pd in my flask…

      1. Martin says:

        I once had to make the Lithium-halide exchange on an electron-rich aryl iodide, and kept getting a product with the same retention time as my starting material (in GC/FID). I kept trying the reaction by changing how long it would react, try the exchange at warmer temperature (-40°C instead of -78°C, etc). In the end, I went and got a GCMS from the lab upstairs from mine (we did not have one), and realized that the product that I had, that had exactly the same retention time as my starting iodide, was the butylated product… What happens is that you do the exchange so fast, and then react the aryllithium with the alkyl halide…
        I ran the reaction at -100°C, worked like a charm!
        I expect in your case, that is what happened as well (with a size of elimination to make the hydrodehalogated product)

    2. Bagger Vance says:

      I’m not technically from Missouri but yeah, i’ll need to see the Org Syn on this before i start placing orders.

  11. Barry says:

    I needed a selective Suzuki coupling onto a heteroaromatic dichloride some years ago. The best conditions we found were biphasic, and gave 8:1 of the desired regio-isomer–as long as you didn’t stir it too well. With efficient stirring, you get 3:1. Seems there were at least two catalytically competent palladium complexes in play, and you didn’t want your substrate to find the less discriminating (more active) one.

  12. Anonymous says:

    1. I recall another example where stirring rates were crucial (and it’s also an aqueous reaction – coincidence?): Org. Synth. 1986, 64, 27. DOI: 10.15227/orgsyn.064.0027 (FREE!) “Some variability of reaction times is probably attributable to differences in stirring efficiency, temperature gradients, and/or particle size of the crystalline starting material.” and “It is a good example of what Turner has called a “point reaction,” as it is very sensitive to experimental details such as temperature and stirring rate. The aqueous buffer procedure given in Part II is a “plateau reaction”.”

    2. I’ve had some reactions that didn’t work well with teflon magnet stirring but do work with mechanical stirring. Some don’t work with teflon paddles but do work with Hershberg (wire) stirrers.

    3. Another aspect of this reminded me of Krapcho decarboxylation with LiCl / DMSO. I think it was an accidental discovery. Further studies showed that it worked with plain DMSO and that a salt wasn’t always needed. I think that stirring rate is a factor, too.

    4. I’m also thinking about multimodal response surfaces (or curves). For instance, with slow stirring you get good result A; with medium stirring you get a worse result B; at faster stirring you get a new better result C. (See also flow rates in liquid chromatography.) Although James Bond currently prefers “Shaken, not stirred,” it might just need to be “Stirred faster, not slower. I need to see that vortex oxygenation!”

    1. bob says:

      A more recent but still bizarre effect of stirring can be found in Baran and Blackmond paper on C-H trifluormethylation. PNAS 108, 14411-14415 (2011) doi:10.1073/pnas.1109059108

  13. DrOcto says:

    Since we’re talking about odd reaction conditions:
    We had a reaction once that was being inhibited by new teflon magnetic stirrers. It was an autocatalytic generation of a silyllithium from R2SiH2 and lithium metal. It turned out that the teflon was quenching the early stages of the reaction.
    Solution: Use an older teflon magnetic stirrer that had been through a similar reaction, no quenching, no problem.

    1. BioChemPhD says:

      What I always wonder is, in such case, how long does it take you to figure it is the stirrer of all things?

      1. John Wayne says:

        Back in the 90’s I worked at Pfizer* and we had a really old, beat up Teflon stirbar we called ‘Igor.’ Some reactions just worked with this thing, and we hypothesized that every catalyst was present on the surface. Now we can add that the surface was also quenched by everything that could quench it.

        * Before they became evil

      2. DrOcto says:

        In this case it didn’t take too long, given that the stirrer bars went into the reaction white, but came out as black as night. It was pretty clear that they were not stable to the R2SiHLi intermediate.

        1. milkskake says:

          teflon surface gest instantly blackened by graphite formation on contact with potassium naphthalenide THF solution. It is a real bummer to spill some of that solution during transfer into Schlenk teflon stpococks, the teflon finger stopcock ceases being air-tight afterwards

  14. Bram says:

    That’s a cool find!
    Seems polarity difference of the solvents is important. With a catalyst very soluble in the organic solvent.
    Interesting to see what would happen with more water soluble Pd cat or substrates

  15. That Nasty Referee says:

    Being a nasty referee as I am, I would have criticized the lack of t-BuLi in the scope.

    1. Reviewer #3 says:

      The SI indicates that t-BuLi was obtained from Albemarle, but curiously it is never mentioned again…

  16. nitpicker says:

    “Most organic chemists picture the quenching of a reactive reagent like an organolithium as pretty much a diffusion-controlled process – the first water atom it sees gets reacted with.”

    Water atom?

    1. Derek Lowe says:

      I was thinking of some hapless H from a water molecule. . .

      1. Jbb says:

        “wasserstoff” 🙂

    2. david says:

      Wasn’t it Peter Beak or Dave Collum who showed some years ago that Br-exchange was diffusion controlled while quenching was not?

  17. houstonchemist says:

    This shows that we know far less than we think we do about alkyllithium reagents even though many of us have used them routinely throughout our careers. They are a black box that suggests further exploration is warranted. This is what I love about organic chemistry. There is always a reaction or reagent that we think we know all to well that results in the occasional scratch you head moment.

  18. Somhairle MacCormick says:

    I love the story here but wonder how they actually came about it. I was thinking maybe they quenched a reaction with brine and then saw the alkyl adduct or something. Can’t believe this was deliberate.

    1. They published a year and a half ago a similar surprising reaction with organometallic compounds. They maybe became expert on looking for surprising coupling in vortex water media …

      1. milkshake says:

        Lipshutz- Krasovsky described aryl zinc Negishi coupling on water about 10 years ago. Aryl zincs are quite tame but they dislike air

  19. Joe Q. says:

    I need to read the paper, but on first glance, this seems like the kind of thing that could be scaled up well with a continuous flow type setup and the right kind of static mixing. So the whole “alkyllithium go boom” issue at larger scales may not be such a barrier.

  20. Not Important says:

    Derek, this is not something new. The Hevia group at Strathclyde have shown these types of conditions used with alkyllithiums and Grignards as far back as 2014.

    1. CrazyChemist says:

      This is a catalytic process! Those you cited are nucleophilic additions (in DESs instead of water and with an excess of organolithium or Grigard reagents).

  21. Gordonjcp says:

    Air? Salt water? Vigorous agitation? I’ve a mental picture of them running it in a sawn-off plastic barrel with an outboard motor on a test stand. Lab equipment by Evinrude… 😉

    1. dave w says:

      That’s about what the equipment for doing it on scale would look like, except for the brand name on the stirring motor.

      1. dave w says:

        Of course, I’d want to carefully evaluate the scaling characteristics of the reaction before trying to run that big a batch…

Comments are closed.