Skip to Content

A Great New C-C Bond Forming Reaction

Is it just me, or do others thing that we’ve been on a run of really interesting new synthetic methods the last few years? I ask partly because of this paper that’s just come out on Angewandte Chemie, from the Baran group. They are, of course, well-known purveyors of new synthetic methods, and this looks like a good one. It’s a decarboxylative coupling between alkyl carboxylic acids (as activated esters, which can be prepared in situ) and aryl boronic acids.

Baran coupling

The organic chemists in the crowd will immediately appreciate a big feature of this system, in that a huge number of of both carboxylic acids and aryl boronic acids are commercially available. These are very common intermediates indeed, so this chemistry opens up a lot of new coupling products. The catalyst is merely nickel chloride, bright green and cheap as dirt, with a phenanthroline ligand added. The paper shows this reaction occurring in the presence of a number of common functional groups, and I’m willing to bet that a lot of synthetic organic and medicinal chemists will look over the paper and see immediate applications for some of their own work.

I also commend the Baran lab for furnishing a 180-page Supplementary Information file, free to download to all comers, that provides extensive details on how to run the reaction. That’s the way to get people to try it! This is especially important here, because it turns out that the reaction, though robust when you get it right, is picky about solvent, base, and concentration (although it can deal with water in the system, and can be run on multigram scale). The only objection I can see someone making is on the grounds of atom economy: the activated ester is a rather large beast, and it looks like it comes back out of the reaction as the pthalimide, so it’s not recyclable as such. For the sorts of chemical diversity that this reaction can provide, though, I think it’s well worth it – and in terms of atom economy, you also have to consider the multistep routes that would have to be used otherwise in many of these cases. This looks to become a standard method, and very soon.

47 comments on “A Great New C-C Bond Forming Reaction”

  1. cp says:

    The FAQ section of the SI is fun.

  2. ABC says:

    One other possible objection, the reaction appears to work only under extremely dilute conditions (0.023 M). It’s great chemistry, but to get any quantity of material you are going to use a lot of solvent.

    1. MTK says:

      0.023M?

      The former process chemist in me threw up a little.

      Full disclosure: I haven’t read the paper or the SI.

      1. choss says:

        well, it’s a proof of concept.

        Surely improving the efficiency of this reaction would be well within the grasp of any red-blooded process chemist.

        1. MTK says:

          We’re chemists, not miracle workers.

          Actually in the SI they did a small concentration study and at double the concentration (0.045M) the yield dropped from 82% to 35%. Another doubling to 0.091M dropped the yield down to 6%. They also looked at temp, solvent, and other things. They didn’t do a DOE which would probably be the way to try and optimize this thing, hoping for some significant multivariable effects.

          1. CMCguy says:

            MTK the use of 1,4 dioxane also would likely be problem for most process chemists to deal with at plant scale, particularly in GMP. Agree at first impression may be suitable for medchem analogs but apparent extensive refinements probably required to implement any leads from this route therefore would require devising a different scheme to make practicable. Isn’t that often the real fun and excitement behind process development?

      2. milkshake says:

        If the reaction is reasonably fast, a reasonable way to emulate pseudo-high dilution conditions on process scale is to slowly pump a concentrated reactant mix and catalyst into solvent preheated to the appropriate temperature.

        I am more concerned here about the need to use 3 equivs of boronic acid, but this is just a first iteration. (The use of dioxane in a free radical-mediated process is also strange, dioxane radical hydrogen abstraction proceeds easily.)

        1. Hap says:

          Didn’t the organozinc coupling require a similar excess of organometallic reagent? Could the excesses be doing something similar in both cases?

        2. Design Monkey says:

          Might be inherent and incorrigible for this type of reactions. After all, it has to catch a not so stable radical, and you better have a sizeable excess for that.

          1. milkshake says:

            They mention occasional formation of arylated dioxane impurities, which would strongly suggest free radical mechanism. Also moisture has detrimental effect (though the reaction does not have to be strictly anhydrous) and the hydrolysis of phthalimidoxy active ester under reaction conditions was observed. The system itself is quite finicky with regards to the narrow range of optimal reaction temperature and reaction mix concentration and acceptable solvents (75 to 85C, dioxane or THF, with added DMF and NEt3), this suggests there are several parts in the catalytic cycle that require just the right combination (in my opinion, it has to do with the reductive pre-catalyst activation in situ, and with the radical chain propagation) and that physical factors like solubility of reactants in the reaction media, the mode of stirring and heating are critical. Maybe the required solvents (dioxane, THF) are not innocuous and their radicals participate in the catalytic cycle. I also think formation of an adduct of boronic acid with tertiary amine is essential for this reaction to work, and that tertiary amine likely acts as a reducing agent for activation of Ni precatalyst. It would be interesting to examine 2-Me and 2,5-diMe THF as a solvent, and to test nonhindered cyclic amines like N-methyl morpholine and DABCO

    2. ravi says:

      We are a CRO and have done more than 100 examples on 1 gram scale !!!! enough said.

    3. ravi says:

      the reaction is easily done on 1 gram scale. i am not entirely convinced that mechanism is radical based. Radical based chemistry requires rigorous exclusion of oxygen and some solvents like dioxane cannot be used.

      i am sure the Baran group will come up with second generation and third generation of these reactions.

      Add this reaction to the medicinal chemist lingo. reductive amination, suzuki coupling, buchwald and now the Baran coupling.

      way to go PSB !!!!!

      just when i thought there was not much more to be discovered in organic chemistry, you come up and show us something so fundamental that no one accomplished.

      PS : I dont like the multiplication (Captcha) that one needs to do in order to post a comment. But I can do the Baran coupling !!

  3. Anon says:

    The FAQ is a work of art. They’ve really made this idiot-proof; an undergrad could do it.

    The solvent-dependence is interesting; only a 10:1 dioxane-DMF mixture worked, and when swapped for THF-DMF the reaction shut down? that’s oddly specific.

    1. Cato the Elder says:

      I would guess b/c of creation of the THF radical

      1. milkshake says:

        but dioxane forms a radical just as easily as THF… What I would like to know is what happens to the extra 2 equivs of boronic acid, the excess that has to be there for a good yield – I think that information would be the key to improving the system

        1. Cato the Elder says:

          Good point… I wonder if the difference lies in the presence of trace inhibitor, as the THF probably came from their solvent system but 1,4-dioxane from a (dry?) bottle that could have had BHT.

        2. Ted says:

          The boronic acid is probably decomposing. I had that problem with aryl boronic acids in Chan-Lam type couplings when the temperature got up over 50°C.

          I was tripped up on the first pilot scaling when we surprisingly hit completion before the pot had finished warming up. Our celebration was short-lived when it turned out our purification wasn’t equipped to handle 2 equivalents of undegraded starting material…

          -t

        3. JFlaviusT says:

          Dioxane has significantly slower H atom abstraction rates than THF because of the inductive effect of the second oxygen. It’s a much larger effect that you might think, which makes dioxane a fine solvent for radical reactions.

  4. anonymous coward says:

    The SI for the activated ester – organozinc paper was also a beast.

  5. MTK says:

    Don’t get me wrong I think this is a great reaction for small scale, but it may not be suitable (yet) for large scale. 20mol% Ni loading is big too. I’d be worried about how to get the residual metal out of the final product.

    How much analytical work in that regard is generally done in discovery before submitting a compound?

  6. triper says:

    MTK they do the reaction on like 10 compounds using 10 mol % as well.

  7. Aspiring chemist says:

    Question for all the med chemists out there-

    How much are these new methods becoming involved in your day to day for making new disconnections or with others at your company in general? By that I mean new nickel methodology (for which a new great method seems to come out weekly by Baran, Weix, etc), but also the multitude of combined nickel-photo catalysis methods being cranked out by Macmillan and Molander or just simple photoredox catalysis.

    Seems like some of these methods are so ridiculously simple and so useful that it would catch on right away without much hesitation.

    1. aspira says:

      see last footnote of the Baran paper. these methods are being copublished with BMS because they are using the chemistry.

  8. Rhenium says:

    I am guessing these examples of huge SI sections were possibly provoked by the “Reproducibility Initiative” from several years ago that investigated a previous Baran group prep that had “issues”.

    The Open Flask blog from the baran group followed not long after, so kudos to them for coming at this so openly. I know they are regular readers here.

    1. triper says:

      It was a Nicolaou group prep.

      1. Phil (not Baran) says:

        Baran was a co-author on that Nicolau group prep, and I would think that made the question of others being able to reproduce his work (and not the work of his students) hit especially close to home. But rather than be defensive and closed off about it (like most people would be), he responded by clarifying the crucial factor which turned out to be the quality of the IBX and the presence of a small amount of water, if memory serves.

        I agree with Rhenium’s observation, and I think Baran’s approach is commendable.

          1. Phil says:

            CJ: Thanks for the clarification, memory only served me halfway. Either way, I was impressed with PSB’s response. He kept it classy (at least in public).

      2. Rhenium says:

        My apologies. Thank you for the correction, see someone else’s reply below.

        Regardless this is an excellent outcome and a very nice paper, just a pity about the dilution conditions necessary.

  9. yuri says:

    I find it very interesting that these rather general, but both atom & catalytically inefficient reactions are getting attention from academic groups and publishers these days. The mindset when I was in grad school would never have considered the pursuit of these type of projects for poor chances of publication in quality journals. It was all about elegance, efficiency, ee and yield vs. practicality for exploring chemical space back then. I am also very happy to see the extensive SI.

    1. eugene says:

      Like yuri, if I had a catalyst that could do something like this at above 5mol%, my boss back in grad school would not consider submitting it anywhere. Unless it was some weird curiosity like activation of alkane CH bonds. I guess I have to change my way of thinking. I still feel kind of funny going to 5 mol% and above catalyst loadings though.

  10. Bro says:

    Hey bro did you see that yield? Maybe it’s static quenched

  11. Blabl says:

    So biologist here with more than passing interest in chemistry (undergrad level), but was interested by the 180 page SI and had a quick browse through that. How did they get away with not reporting SEMs on any of their numbers in the SI?! Isn’t this supposed to be a well respected journal? In biology you’d get crucified if you’d try to do that.

    1. Phil says:

      This made me laugh, not at you, but at the general statistical illiteracy of synthetic organic chemists (and I was one of them until I got into industry and started learning stats on my own). Standard errors are not reported because most of the experiments are conducted only once.

      To be fair, within the organic chemistry community, yields are generally acknowledged as “fuzzy,” especially when many experiments are run on such small scale that user error can change the resultant yield a lot (i.e. theoretical yield of 5mg, losing 0.5 mg gives a 10% swing in yield). No experienced organic chemist would look at a table with reported yields of 70%, 72%, 75% and 77% and think these reactions were distinguishable from one another in terms of efficiency. It’s more like ranges: poor (0-25%), not great (25-60%), serviceable (60-80%), good (80-90%), very good (90+%). These are my ranges for judging academic results, other people will have their own.

      Taking a reaction that is at least 80% yield in a paper and optimizing it to be 95% yield+ is generally feasible, below that it gets less and less likely. It’s only during this course of optimization that organic chemists start applying statistical analysis (DoE, full factorial modeling, etc).

  12. Barry says:

    at higher concentration, I’d expect simple Kolbe-like dimerizations

    1. milkshake says:

      This reaction presumably depends on Ni(I) species dumping one electron into electron deficient tetrachlorophthalimidoxy ester. The produced RCO2(.) radical decarboxylates, R(.) is trapped either by boronic acid or more likely by Ni(II) species. And so on. It should be possible to control how much Ni(I) is around by varying the reducing agent, which in this case is most likely the triethylamine, although boronic acid-amine complex is also a good candidate.

  13. Mikkel says:

    After reading this blog for quite some time, I do wonder why Barans work is mentioned so often? It just makes you wonder, and I am not even a competitor in the field.
    It is an interesting cross coupling, but very exotic and almost non-cat (20 mol % Ni cat…), and to me, as a process chemist, completely useless. I do suspect that in most situations good alternatives (routes) exist. You can always find exceptions. Since synthetic organic chemistry is such a mature science (by some standards), I wish people would develope more sustainable (green if you like) and in the broad sense more scalable chemistry (reliable yields…) without the “10- 20 mol % catalyst” and advanced ligands and/or reagents.
    There more than 90 naturally ocurring elements, we dont need to try every combination.

    1. kriggy says:

      “There more than 90 naturally ocurring elements, we dont need to try every combination.”

      So you are suggesting that chemists stop developing new reaction in favour of optimizing the “old reactions”? Also if ppl were following your opinion, there would be no suzuki coupling or similar reactions and I am pretty sure those are used in industrial setting. Maybe this reaction is not catalytic and is not suitable for you as a process chemist but maybe someone can make it catalytic in the future

    2. Derek Lowe says:

      This is definitely not a process reaction (yet), but I can assure you that the med-chem folks over in the other hallway will get use out of this one. We’re after fast molecular diversity, and this reaction definitely delivers that.

  14. Mikkel says:

    Kriggy, no I am not, but I do see many reactions popping up, initially being hailed as fantastic, then slowly fading away. I dont think we should compare it with the Suzuki coupling. Time will tell.
    Technically it is catalytic, just not a lot.
    I actually stole that “there are more than 90…” sentence from Barry Trost, his introduction in a speech a few years ago. Nature would agree.
    Derek, I doubt this reaction will ever be a process reaction, process people tend to dis-like nickel salts in large amounts. But perhaps someone finds a different more benign catalyst in time. Interestingly, Baran once told our people (as a consultant) that he reads OPRD “religiously”. Perhaps he will put more science into it than faith in the future.

    1. Kriggy says:

      I agree with you but I think this one has potential. It realy depends on what you are doing and I think if you are making drug where one dose is in range of single miligrams where single batch is maybe 500 g (and pretty expensive too) I dont think the 20 cat. % realy matters because the profit is so huge that it easily covers the expensive catalyst or its removal. Of course, its not realy feasible to run it on tonne scale.
      Anyway, Im wondering if the catalyst could be immobilized on solid support and the reaction run in flow reactor?

      1. Albert says:

        In my opinion bigger issues which preclude this reaction from being process friendly (>100 kg scale) is not the nickel loading which could probably be decreased by more screening, but expensive and not easily recoverable activating group as well as high excess of boronic acid. Of course low concentration is a killer too if indeed there is no way to increase it.

      2. Phil says:

        “Anyway, Im wondering if the catalyst could be immobilized on solid support and the reaction run in flow reactor?”

        http://onlinelibrary.wiley.com/doi/10.1002/adsc.201500846/abstract

        1. Kriggy says:

          Thank you 🙂

  15. Albert says:

    Another process chemist here. I understand completely the utility of this kind of stuff in the med-chem context, but I think Mikkel is also right about this being an unlikely candidate for a process even with a lot of development. Too many shortcomings and in most contexts one could easily devise alternative routes for such a disconnection.

  16. chemstud22 says:

    I do agree that we have been on a run of really interesting new synthetic methods, especially in regard to the numerous alkyl/aryl cross-couplings in the literature recently. But I also agree with some of your readers in that this work from Baran, while nice and useful, is nothing exceptional and is an elegant extension of his past recent work. If we’re going to be discussing “great new C-C bond” forming reactions I feel it’s our duty to analyze/compare and contrast popular methods in the field and wait to see who takes the cake (metaphorically speaking) years from now. There are exciting strategies being unveiled very frequently so I think we should be using a wider lens in these discussions. For example, it would be worthwhile if chemists could get insight and feedback into the utility and reliability of the alkylation reactions developed by Baran, Molander, MacMillan etc. (ie. MacMillan’s recent JACS employed 0.5 mol% Ni and Molander’s alkyl silicates are extremely cheap)

Leave a Reply

Your email address will not be published. Required fields are marked *

Solve the math problem. * Time limit is exhausted. Please reload CAPTCHA.